Forensic Dna Evidence For Criminal Justice Professionals: Introduction To [PDF]

Citation preview

Introduction to



FORENSIC DNA EVIDENCE FOR CRIMINAL JUSTICE PROFESSIONALS



Introduction to



FORENSIC DNA EVIDENCE FOR CRIMINAL JUSTICE PROFESSIONALS Jane Moira Taupin



Boca Raton London New York



CRC Press is an imprint of the Taylor & Francis Group, an informa business



CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2014 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Version Date: 20130503 International Standard Book Number-13: 978-1-4398-9910-6 (eBook - PDF) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www. copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com



Contents Preface..................................................................................................................xi Acknowledgments............................................................................................xv About the author............................................................................................ xvii Chapter 1 History of forensic DNA profiling in criminal investigations................................................................................ 1 1.1 Discovery of structure and importance of DNA molecule: A Nobel prize................................................................................................ 1 1.1.1 Mendelian law of inheritance................................................... 1 1.1.2 Structure of DNA........................................................................ 2 1.1.3 Human genome project............................................................. 3 1.2 DNA and concept of individuality........................................................ 4 1.3 Alec Jeffreys and the world’s first murder case solved by DNA....... 5 1.4 Early criminal court challenges to DNA technology.......................... 7 1.5 Changing the face of forensic science: The value of biological evidence..................................................................................................... 9 References.......................................................................................................... 11 Chapter 2 Strengths and limitations of DNA profiling evidence...... 13 2.1 Introduction: Power and caution.......................................................... 13 2.2 Discrimination power of DNA profiling............................................. 15 2.3 Genetic basis for DNA profiling........................................................... 16 2.4 Stability of DNA profiling..................................................................... 18 2.5 Persuasive statistics................................................................................ 19 2.6 Relatives................................................................................................... 21 2.7 DNA databases....................................................................................... 22 2.8 DNA ­intelligence-­led policing.............................................................. 23 2.9 Mass disasters......................................................................................... 24 2.10 DNA evidence in context....................................................................... 24 2.11 Time of deposition: Transfer and persistence of DNA...................... 25



v



vi



Contents



2.12 Relevant evidence?................................................................................. 27 2.13 Relevant exhibits?................................................................................... 29 2.14 “CSI effect” and the notion of infallible forensic evidence.............. 30 2.15 Relationships of lawyers and scientists............................................... 30 References.......................................................................................................... 30 Chapter 3 DNA profiling basics................................................................ 33 3.1 What is DNA?.......................................................................................... 33 3.2 Biological materials allowing DNA profiling..................................... 33 3.2.1 Searching for DNA on exhibits.............................................. 35 3.2.2 Blood........................................................................................... 37 3.2.3 Semen and spermatozoa.......................................................... 38 3.2.4 Saliva.......................................................................................... 39 3.2.5 Hair roots................................................................................... 40 3.2.6 Dandruff and skin.................................................................... 40 3.2.7 Nasal secretions........................................................................ 40 3.2.8 Vaginal secretions..................................................................... 40 3.2.9 Sweat........................................................................................... 41 3.2.10 Wearer DNA.............................................................................. 41 3.2.11 Touch DNA................................................................................ 42 3.2.12 Urine and feces......................................................................... 43 3.2.13 Emerging techniques............................................................... 43 3.3 Reference samples.................................................................................. 44 3.3.1 Buccal scrapes........................................................................... 44 3.3.2 Blood........................................................................................... 44 3.3.3 Plucked hair samples............................................................... 44 3.3.4 Personal belongings................................................................. 45 3.4 Current profiling technique: Short tandem repeats (STRs).............. 45 3.5 Reading tables of alleles........................................................................ 47 3.6 Obtaining DNA profiles........................................................................ 51 3.6.1 Controls...................................................................................... 51 3.6.2 Extraction................................................................................... 52 3.6.3 Quantification........................................................................... 52 3.6.4 Amplification............................................................................. 53 3.6.5 Separation and detection......................................................... 54 3.6.6 Reading electropherograms.................................................... 54 3.6.7 Artifacts and other technical issues....................................... 55 3.7 Time required to obtain DNA profiles................................................ 59 3.8 Designating peaks.................................................................................. 60 3.9 Case documentation and review.......................................................... 61 References.......................................................................................................... 63



Contents



vii



Chapter 4 Evidential value and statistics................................................. 65 4.1 Introduction............................................................................................. 65 4.2 Interpreting DNA profiles..................................................................... 66 4.3 Statistical approaches and obtaining final statistics......................... 66 4.3.1 Random match probability and likelihood ratio................. 66 4.3.2 Calculating frequencies........................................................... 68 4.3.3 Comparison of probability of exclusion and LR methods.... 69 4.3.4 Identity and rarity.................................................................... 71 4.4 Legal fallacies.......................................................................................... 71 4.5 Understanding reports: Common phrases and their meanings...... 73 4.5.1 Inclusion and exclusion........................................................... 73 4.5.2 Declared contributor................................................................ 75 4.5.3 Verbal descriptors..................................................................... 76 4.6 Sampling correction and uncertainty.................................................. 76 4.7 Relevant population and impact on statistical value........................ 77 4.8 Relatives................................................................................................... 77 References.......................................................................................................... 78 Chapter 5 Partial profiles, low levels, and mixtures............................. 81 5.1 Partial profiles......................................................................................... 81 5.2 Low level and suboptimal profiles....................................................... 82 5.2.1 Definitions................................................................................. 82 5.2.2 Stochastic effects....................................................................... 85 5.2.3 An interesting experiment...................................................... 88 5.2.4 Enhancement techniques........................................................ 90 5.2.5 Improving reliability of results.............................................. 90 5.2.5.1 Biological (consensus) model................................. 91 5.2.5.2 Statistical (probabilistic) model.............................. 91 5.2.6 Contamination.......................................................................... 92 5.3 DNA mixtures from two or more people........................................... 93 5.4 Mixture interpretation steps................................................................. 98 5.5 Low template mixtures.......................................................................... 99 5.6 Complex mixtures................................................................................ 100 References........................................................................................................ 101 Chapter 6 Y-STR profiling........................................................................ 105 6.1 Introduction........................................................................................... 105 6.2 Benefits................................................................................................... 106 6.3 Theory.................................................................................................... 109 6.4 Statistics..................................................................................................110 6.4.1 Frequency estimates of Y ­ -­STR haplotypes...........................110 6.4.2 Meaning of ­Y-­STR match........................................................111



viii



Contents



6.5 Number of male contributors to ­Y-­STR profile.................................113 6.6 Determining mixture ratios.................................................................113 6.7 Combining statistics from autosomal and ­Y-­STR profiling............114 References.........................................................................................................114 Chapter 7 Other DNA techniques including mitochondrial DNA.... 117 7.1 Introduction............................................................................................117 7.2 DNA analysis of bone...........................................................................117 7.3 Mitochondrial DNA basics...................................................................119 7.4 Statistics in mitochondrial DNA analysis......................................... 122 7.5 Contamination...................................................................................... 124 7.6 Mixture mitochondrial DNA profiles................................................ 125 7.7 Familial DNA searching...................................................................... 126 7.8 Domestic animal hair........................................................................... 128 7.9 Other techniques.................................................................................. 129 References........................................................................................................ 129 Chapter 8 Concerns and controversies................................................... 131 8.1 Introduction........................................................................................... 131 8.2 Quality issues........................................................................................ 132 8.3 Relevant sample testing....................................................................... 133 8.4 Contamination...................................................................................... 134 8.5 Interpretation issues............................................................................. 138 8.6 Error rates.............................................................................................. 138 8.7 Overreliance on DNA technology...................................................... 140 8.8 Interpretation of DNA profiles: Objectivity and subjectivity.........141 8.9 ­Retesting of samples............................................................................. 143 8.10 Adversarial system............................................................................... 144 8.11 Misconception about exact science.................................................... 144 8.12 Obligations............................................................................................. 144 References........................................................................................................ 145 Chapter 9 DNA pointers for criminal justice professionals.............. 147 9.1 Introduction........................................................................................... 147 9.2 Advantages of DNA profiling............................................................. 148 9.3 Querying DNA evidence: Advice for the prosecution and the defense................................................................................................... 148 9.4 Warning signs....................................................................................... 149 9.5 Was all evidence tested?...................................................................... 150 9.6 Pretrial review....................................................................................... 151



Contents 9.7



9.8



ix



Suggested cross-examination questions........................................... 151 9.7.1 General..................................................................................... 151 9.7.2 Single source DNA profiles associated with blood, semen, or saliva....................................................................... 152 9.7.3 Difficult DNA profiles (partial, low level, mixture, unspecified origin)................................................................. 152 9.7.4 Expert witness......................................................................... 153 Discovery requests............................................................................... 153



Appendix A: Glossary of terms used in reports and testimony.............. 155 Appendix B: Selected DNA issues and case examples...............................161 Appendix C: Steps in review of evidence.................................................... 163 Index................................................................................................................. 165



Preface Purpose of this book This book aims to provide trial lawyers and criminal justice professionals with sufficient tools to understand and probe the evidential value of forensic DNA evidence in their criminal cases. This text is designed for criminal lawyers and criminal justice professionals and written for ­­nonscientific readers. The criminal justice professional will gain specific knowledge of the strengths and limitations of DNA evidence in criminal cases. The prosecution lawyer will improve his or her understanding of DNA evidence— when such evidence should be emphasized, when to discuss the evidence with forensic experts, and when to proceed with trial despite a lack of DNA evidence. The defense lawyer may be better equipped to challenge DNA evidence and perhaps employ an independent expert, understand when to focus on other aspects of the prosecution’s case, and know when to secure the advantage of an early guilty plea. The extreme probabilities quoted in many criminal cases with DNA evidence may make it appear that such evidence allows no margin for error. Many lawyers do not even know what “one in one trillion” means, yet this number is cited in many forensic reports. However, recent cases worldwide have shown that DNA evidence is not infallible. There is a danger in relying on DNA statistical probabilities in the determination of guilt. DNA evidence is just one piece of circumstantial evidence and does not prove an accused was the offender. Also, DNA profiling is performed and interpreted by humans and thus, like all scientific testing, is subject to quality control and other errors. DNA profiling for use in forensic cases was a significant scientific achievement in the 1980s and has been considered the most innovative technique in forensic science since fingerprinting. Yet how can a lawyer with limited scientific knowledge grasp the technology and exploit it or challenge it? This text aims to provide criminal lawyers with the xi



xii



Preface



knowledge to do just that and make DNA profiling less intimidating. DNA profiling should not be a “Pandora’s box” that both prosecution and defense attorneys are afraid to open for fear of what may be unleashed. All sides of the adversarial system should be confident and aware of the strengths and limitations of DNA evidence in order for the criminal justice system to operate effectively. This book aims to assist in this process so that a criminal lawyer can confidently understand and convey this innovative technology in a courtroom. DNA offers a degree of certainty often missing in a criminal trial. It helps focus police investigations and solve ­­ decades-­ old cold cases. Alternatively, the investigating police may be happy with a DNA result and become complacent, and thus fail to explore or reject other avenues of evidence. Counsel for both parties should be aware of biased or sloppy investigations. Similarly, forensic scientists, judges, and jury members should be aware of the limitations and strengths of DNA evidence. This book aims to provide readers with the tools to recognize both aspects of testing. The “CSI effect” may impact the judge, jury, and even legal practitioners so that the DNA evidence is accorded more weight than it warrants. Again, this text hopes to provide a balanced perspective on the weight of DNA evidence in criminal trials. The understanding of DNA concepts by the jury is also an issue. Ultimately, it is up to legal counsel to convey the concepts of DNA profiling to a jury in a manner that is readily understood. This text is intended to provide the tools for criminal lawyers to use DNA profiling effectively.



Scope and limitations This book is designed for lawyers and other criminal justice professionals (such as legal researchers) who have limited scientific knowledge. However, forensic scientists and those interested in the application of DNA to criminal proceedings should also find it valuable. It should be thought-­ ­­ provoking reading for crime authors, journalists, and legal commentators. Most of the background scientific information needed to understand the basis of DNA profiling is provided in this text. Although some concepts may appear too complex or scientific on initial perusal, readers are encouraged to grasp the concepts with which they are comfortable and engage in further reading and/­or discussions among colleagues. Literature references to scientific papers are provided but it is not necessary for readers to peruse any or all of them. However, readers should be aware of the references because of their potential to be mentioned in court. A forensic scientist providing DNA evidence at trial should be familiar with some, if not most, of the literature quoted, especially the most



Preface



xiii



famous papers cited in this book. The references are intended to assist trial lawyers in their examinations of expert witnesses. Understanding and quoting the literature will enable them to assess an expert witness’s knowledge of the matters discussed therein. The text covers the most common DNA methods used in criminal trials today—nuclear DNA short tandem repeat (STR) techniques, mitochondrial DNA, and ­­Y-­STR profiling. The statistics discussed are the probabilities obtained by comparisons of reference samples and crime scene samples. Paternity statistics will not be explored. The book is applicable to the adversarial (trial and jury) legal systems encountered in countries such as England, the United States, and Australia. Many of the principles can be applied in other countries that do not utilize juries but use the inquisitorial approach (courts in Europe consist of panels of judges and a single judge presides over court proceedings in Middle Eastern countries). Since this text does not discuss the forensic principles in depth or assume scientific knowledge on the part of readers, it contains numerous references to specific forensic science and statistical texts that can provide more complete discussions on particular topics. There are also references throughout the book to case studies from around the world that illustrate one or more particular applications of DNA profiling evidence to criminal proceedings and whether that evidence has been applied effectively. The intent is to provide legal professionals with pointers that may be applicable to their own cases. The final chapter includes a handy list of questions for a criminal justice professional to consider when trying a case involving the use of DNA evidence.



Acknowledgments I thank Becky Masterman, senior editor at Taylor & Francis Group for her support and enthusiasm for this topic and, as always, her endless patience. I thank my former colleagues at LGC Forensics England for their advice on DNA profiling when I moved there from Australia. Special thanks go to Pauline Stevens and Craig Davies, LGC’s “DNA gurus.” I also thank Craig for his continuing advice in the advancing field of DNA profiling evidence. Thanks also go to Dr.  Roland van Oorschot at the Victoria Police Forensic Services Department in Melbourne for his support while I was at this laboratory, especially for his help in publishing case studies. His own continuing flood of publications in the DNA literature shows his enthusiasm for his specialty. I appreciate the suggestions of Victoria barrister Peter Chadwick, S.C. indicating what a barrister would like to see in a DNA book. I also thank the many barristers and lawyers with whom I worked, who devoted considerable time and effort to understanding a topic that is often alien to legal personnel. Special thanks go to New South Wales Public Defenders Ian Nash and Richard Wilson, and Victoria barristers Alan Hands, Benjamin Lindner, Moya O’Brien, Carmen Randazzo, and Samantha Poulter. This book is for all the criminal justice professionals who are confronted with DNA reports and may have little time to understand them or prepare cases involving DNA evidence. I hope you find it useful.



xv



About the author Jane Moira Taupin obtained a BS (Honours) from the University of Melbourne in Australia. After graduating, she accepted research positions at the University of Melbourne, first in antibody production at the Howard Florey Institute and then in cancer research at the Austin Hospital. She joined the Australian Federal Police as a Constable and advanced to Stage 1 Detective, working in diverse areas including drug surveillance and government fraud. She was transferred temporarily to the only atomic energy facility in the country (Lucas Heights) and used neutron activation analysis on a number of criminal cases. She left to join the Victoria Police Forensic Services Centre as a forensic scientist where she pursued a wide variety of major crime cases involving biological evidence. Taupin investigated crime scenes for blood pattern analysis and conducted searches for biological fluids. She has presented biological expert evidence in courts of law since 1987 and has presented DNA profiling evidence in court since 1999. Taupin earned a postgraduate diploma along with an MA, both in criminology from the University of Melbourne. Her master’s thesis in 1994 on the impact of DNA profiling was one of the first in the field. She then moved to Forensic Alliance in England where she performed similar work in the company’s Oxford and Manchester laboratories. When LGC Forensics took over the company, she became a lead scientist. In December 2009, she returned to MRS Limited in Melbourne as an international forensic auditor. She also lectured in Qatar and Bahrain on a variety of subjects including DNA analysis. She is currently an independent forensic consultant and trainer. Taupin has published many articles in p ­ eer-­reviewed journals discussing trace evidence, clothing damage, and blood pattern analysis. She also ­co-­authored a text on the forensic examination of clothing. She won a Young Investigator’s Award from the International Association of Forensic Sciences and attended its meeting in Tokyo in 1996 in recognition of her work on clothing damage analysis. The following year she won a Michael Duffy travel fellowship from the Australian xvii



xviii



About the author



government to attend the American Academy of Forensic Sciences meeting in New York and visit international laboratories including the FBI in the United States, the Forensic Science Service in England, and the BKA in Germany. She participated on the inaugural committee of the Scientific Working Group on Hair (SWGHAIR) under the auspices of the FBI in Washington for 6 years. In 2009, she received a Good Citizen Award from the Greater Manchester Police in England for her work in helping to solve a horrific case of rape of an elderly woman through DNA profiling evidence.



chapter one



History of forensic DNA profiling in criminal investigations Much of the history of deoxyribonucleic acid (DNA) is truly fascinating and as compelling as any b ­ est-­selling crime novel. Scientists, particularly those in the biology discipline, are familiar with the breakthroughs that led to the use of DNA in criminal work. They use these stories as building blocks to assess new information; the criminal justice professional may benefit from these stories as well. This chapter will briefly introduce the legal professional to some important discoveries and famous scientists who unlocked the secret of how life may work and how the key was applied to forensic casework. Scientific terminology used will be more fully explained in later chapters.



1.1 Discovery of structure and importance of DNA molecule: A Nobel prize 1.1.1 Mendelian law of inheritance The history of DNA can be thought of as beginning with an Austrian (now Czech Republic) Augustinian monk, Gregor Mendel, who is called the founder of genetics. In 1865, he completed a series of experiments with peas and showed that certain traits, such as shape and color, were inherited in different packages—now called genes. When crossing white flower and purple flower plants, Mendel found that the result was a purple flower, not a blend of the colors. He conceived the idea of heredity units that were either recessive or dominant (Mendel, 1865). These units (genes) normally occur in pairs in body cells but segregate during the formation of sex cells. A dominant gene will hide a recessive gene. Mendel stated that each person inherits two traits from each parent. If the traits are the same, they are homozygous; if they are different, they are heterozygous. This becomes an important factor for forensic scientists and lawyers when interpreting DNA profiles. The alternative forms of each trait are called alleles. The two principles Mendel described were the Law of Segregation, by which each parent passes a randomly selected allele to its offspring 1



2



Introduction to forensic DNA evidence for criminal justice professionals



during fertilization, and the Law of Independent Assortment, by which separate genes for separate traits are passed independently from parent to child. Of the 46 chromosomes in a human cell, half are from the mother’s egg and the other half from the father’s sperm. The sex chromosomes X and Y are incorporated into every DNA profile.



1.1.2 Structure of DNA Until a scientific paper was published in 1944 (Avery et al.), biologists thought that genes—the units of inheritance—were made of proteins. Oswald Avery, an American scientist, managed to transfer the ability to cause disease from one strain of bacteria to another, showing the connection between nucleic acids and genes. This paper has been described by Nature, a leading scientific journal, as the defining moment in nucleic acid research. The following decade saw the amazing discoveries of the structure of DNA and how it is copied from one generation to the next. Linus Pauling in California postulated a triple helical structure for DNA in 1953. So had James Watson and Francis Crick working at Cambridge in England, but they were all wrong. It was Rosalind Franklin’s x-ray diffraction photograph that was (unbenownst to her) shown to James Watson that revealed the true structure of DNA to Watson and Crick. Nature considers the year 1953 an annus mirabilis (year of wonders) for science. The ­three-­dimensional structure of DNA was first described by Watson and Crick in April 1953 in a Nature article. This was the first explanation of how genetic information is encoded and transferred from one generation to the next. This classic paper first describes the double helical structure of DNA. Nature later stated that the authors noted “with some understatement that the structure suggests a possible copying mechanism for the genetic material.” Another paper in the same issue of Nature analyzes the x ­ -­ray crystallography evidence and suggests that a double helical structure exists in biological systems (Wilkins et al., 1953). Following on from this, Rosalind Franklin and Ray Gosling, her student, provided further evidence of the helical natures of nucleic acids and concluded that their phosphate backbones lie on the outsides of the structures (Franklin and Gosling, 1953). In the next month’s issue of Nature, Watson and Crick followed up with largely accurate speculation on how the base pairing in the double helix allows replication of DNA (Watson and Crick, 1953). Watson and Crick and Maurice Wilkins were awarded the Nobel Prize in Physiology and Medicine in 1962 for their discovery of the double helix structure of the DNA molecule (Figure 1.1). A b ­ est-­selling memoir by James Watson (1968) gives his account of the efforts to beat Linus Pauling



Chapter one:  History of forensic DNA profiling in criminal investigations



3



Base pairs Adenine



Thymine



Guanine



Cytosine



Sugar phosphate backbone



U.S. National Library of Medicine



Figure 1.1  Structure of DNA. The two strands of the double helix are connected by base pairs. The bases are adenine and thymine, and guanine and cytosine. (Source: U.S. National Library of Medicine.)



to solving the structure of “life.” Watson was only 24 when he made the discovery and the book portrays scientists as eminently human, with petty rivalries and driving ambitions. Many people today believe that Rosalind Franklin also should have been awarded the Nobel Prize but she was never even nominated, and died of cancer at age 37 in 1958. Watson, in an epilogue to his memoir, acknowledges his incomplete and unjust depiction of Franklin. He later helped establish the Human Genome Project (see next section).



1.1.3 Human genome project One of the misconceptions that the public may carry is that DNA profiling examines all areas of the DNA molecule. This is currently impossible to do for every criminal case. In fact, the entire set of genes in DNA had never been identified until long after the Human Genome Project commenced in 1980. The project’s goal was to identify the 20,000 or more genes and determine the sequences of the 3 billion base pairs that make up human DNA. A genome consists of the entire DNA in an organism including its genes.



4



Introduction to forensic DNA evidence for criminal justice professionals



After 13 years of laborious and meticulous work, the Human Genome Project was completed in 1993. It was coordinated by the U.S. Department of Energy and National Institutes of Health and the Wellcome Trust in the U.K. Additional contributions came from Australia, Japan, France, Germany, China, and other nations. It is possible that when DNA sequencing techniques progress further, direct comparisons of very large DNA segments may be done. It is even conceivable that comparisons of whole genomes may be feasible and thus allow precise individual identification in the future. At the moment, however, we are reliant on the observations of a designated number of areas on the DNA molecule and statistical analysis in the comparison between the individual and an evidence sample. Furthermore, each individual’s DNA is approximately 99.9% the same as another’s, so that we are concerned with only very small differences in their sequences.



1.2 DNA and concept of individuality Today most people understand that DNA makes every individual unique—except for identical (monozygotic) twins who emerge from the same egg. The use of the term DNA profiling without a prefix generally implies autosomal STR profiling. This is the analysis of the DNA in the nucleus of the cell from chromosome to chromosome (autosomal relates to chromosomes) and is the most common DNA profiling method used in criminal cases. The term STR is discussed in Chapter 2. Y-STR profiling is a specific kind of nuclear DNA profiling which analyzes the Y chromosome only, is inherited paternally and is much less discriminatory (discussed in Chapter 6). Mitochondrial DNA (mtDNA) profiling analyzes the DNA in mitochondria located outside the nucleus of the cell, is also less discriminatory and is inherited maternally (discussed in Chapter 7). DNA is a complex chemical considered to be a genetic blueprint that determines our chemical and physical characteristics. The genes carry information for making all the proteins required by an organism. These proteins determine how an organism looks, how well it fights infection, and sometimes how it behaves. Other sequences of DNA have structural purposes or are involved in regulating the use of this genetic information. As described above, half of human DNA is inherited from the mother (from the egg that is fertilized) and half from the father (from spermatozoa). Forensic DNA profiling examines locations along the DNA molecule that are highly variable from one individual to another. These regions are considered junk DNA segments that appear to have no known function in the human body. DNA is a nucleic acid consisting of two long polymers called nucleo­ tides, with backbones made of sugars and phosphate groups. The two



Chapter one:  History of forensic DNA profiling in criminal investigations



5



polymer strands run in opposite directions and form the shape of the double helix as first described by Watson and Crick. Attached to each sugar molecule is one of four similar chemicals called bases. These bases are called adenine, thymine, guanine, and cytosine and are repeated billions of times throughout a genome. The particular order of the bases is important and is responsible for life’s diversity. These bases also become important in the interpretations of DNA profiles in criminal cases. DNA within a cell is organized into long structures called chromosomes. These chromosomes are duplicated before cells divide in a process called DNA replication. It is naturally accepted in recently published papers that the main precondition for the transfer of genetic information from one cell to its daughter cells and from one generation to the next is the high stability of DNA (Vennemann and Koppelkamm, 2010).



1.3 Alec Jeffreys and the world’s first murder case solved by DNA When this author commenced work as a forensic scientist in 1986, the concept of DNA profiling in criminal cases was unknown. Conventional genetic markers such as ABO blood grouping and enzyme typing were used on physiological fluid stains such as blood and semen. These markers used the theories of population genetics that DNA analysis also uses; however these markers have very low discrimination power. For example, the blood group O is common to nearly half the population. Often due to the ages and sizes of stains on clothing or weapons, limited or no information is obtained due to the instability of the enzymes and blood groupings and the susceptibility to bacterial attack. It was frustrating to forensic scientists to find noticeable quantities of blood and semen on exhibits from crime scenes and obtain no evidential value from them. These early serology tests have now been superseded in most forensic laboratories. DNA profiling was introduced into the forensic arena by notable publications from English scientists, again in Nature (Jeffreys et al., 1985; Gill et al., 1985). Professor (now Sir) Alec Jeffreys of Leicester University discovered in 1984 that VNTRs (variable numbers of tandem repeats) were present in all human DNA but varied in length for each individual. He called his method DNA fingerprinting because it typed many “minisatellites” simultaneously so that the final result resembled a distinct barcode. This technique, using restriction fragment length polymorphism (RFLP), was used first in a disputed immigration case to confirm the identity of a British boy whose family was originally from Ghana (Jeffreys et al., 1986). The case was resolved when DNA results showed the boy was closely related to other members of the family. The first murder case



6



Introduction to forensic DNA evidence for criminal justice professionals



solved by DNA profiling (Case 1) was prosecuted in the English midlands in the mid 1980s (Gill and Werrett, 1987; Wambaugh, 1989). Case 1 Lynda Mann, a 15-year-old girl, was found raped and murdered in 1983. Three years later, 15-year-old Dawn Ashworth was raped and murdered nearby. A kitchen porter confessed to the second murder but not the first; police, however, were convinced both girls had been murdered by the same offender. The semen on both bodies fell into the type A blood group and the enzyme profile matched 10% of the adult male population. The police asked Professor Alec Jeffreys to analyze the samples using his new technique of DNA fingerprinting. The semen on both bodies was indeed from the same man but DNA profiling excluded the kitchen porter. This led the police to conduct a world first—a DNA screen of 3,000 local men. Colin Pitchfork persuaded a work colleague to donate a sample for him, but police discovered the ruse and subsequently found that the DNA profile of the semen on the bodies matched Pitchfork’s DNA profile. Pitchfork was convicted and sentenced for the two murders in 1988. Professor Jeffreys was only 34 years old at the time, a young genius like James Watson. His work on the murder case was confirmed by Dr. Peter Gill and Dr. Dave Werrett of the Forensic Science Service (FSS) in England, who jointly published the first paper on applying DNA profiling to forensic science (Gill et al., 1985). The first suspect in the Pitchfork case was thus the first to be exonerated by DNA profiling. The high discrimination power—the power to exclude—arguably produced the greatest impact of DNA profiling in the criminal justice field. The Pitchfork case led to the rapid introduction of DNA analysis into casework in England and Wales (Werrett et al., 1989). After the polymerase chain reaction (PCR) technology was developed in the late 1980s, DNA typing methods began incorporating PCR and RFLP technology was phased out. PCR is essentially a molecular photocopier that can amplify very small samples into quantities that can be detected and analyzed. This method is a boon for forensic science as it enables the analysis of minute quantities of blood and semen, as well as degraded samples such as those commonly encountered at crime scenes. Today, microsatellites are used instead of minisatellites.



Chapter one:  History of forensic DNA profiling in criminal investigations



7



Professor Jeffreys first exploited DNA analysis to confirm the identity of Josef Mengele (Jeffreys et al., 1992), the notorious Nazi concentration camp physician who performed horrific experiments on Auschwitz inmates. The DNA from an exhumed femur bone of the presumed skeleton of Mengele (buried in Brazil as Wolfgang Gerhard) was compared to the DNA from his widow and son. The DNA fingerprinting term has been dropped in favor of DNA profiling to eliminate the poor analogy with fingerprint technology, which has no genetic statistical basis. Chapter  3 discusses the current profiling techniques.



1.4 Early criminal court challenges to DNA technology One of the first court cases in the United States using DNA profiling was conducted in November 1987 in Orange County, Florida. Tommy Lee Andrews was convicted of rape based on testing that matched DNA from his blood with DNA from semen taken from the victim (Aronson, 2007). The prosecution merely had to prove the reliability of the evidence, not general scientific acceptance. Aronson’s book describes how quality assurance and control were lacking in the early stages of DNA profiling and how subsequent defense challenges improved quality at all stages of the process including collection and sampling. The West Virginia Supreme Court was the first state high court to rule on the admissibility of DNA evidence in the case of State v. Woodall in 1987. The court accepted DNA testing by the defendant but inconclusive results failed to exculpate Woodall. The court upheld the conviction for rape, kidnapping, and robbery of two women but Woodall continued to pursue further testing. DNA analysis using PCR determined that he was innocent and he was released from prison in 1992. The real perpetrator was eventually found (Innocence Project). The first case that seriously challenged the admissibility of DNA evidence was People v. Castro heard in the New York Supreme Court. Castro was accused of murdering his neighbor and her 2-year-old daughter (National Research Council, 1992). Blood found on Castro’s watch contained DNA that matched the DNA of the dead woman but Castro claimed the blood was his own. During a pretrial hearing, the court held that DNA was generally accepted in the scientific community but that the technique as applied in the Castro case was so flawed that the evidence of a match was not admissible, although evidence indicating that the DNA was not Castro’s was admissible. The court concluded that pretrial hearings are



8



Introduction to forensic DNA evidence for criminal justice professionals



required to determine whether a testing laboratory’s methods meet scientific standards and are reliable. The defendant subsequently pleaded guilty. As a result of this case, the scientific and legal communities raised concerns about DNA profiling in the United States. Pretrial hearings may also be held in England to decide the admissibility of DNA evidence in question. In the case of R. v. Doheny and Adams (1997), the court noted that the risk of laboratory error, the method of DNA analysis used, and the basis of statistical calculations should be examined. The O.J. Simpson trial (Case 2) focused the attention of the American public on forensic evidence and DNA in particular. Case 2 Former American football star and actor O.J. Simpson was tried on two counts of murder following the June 1994 deaths of his ­ex-­wife Nicole Simpson and her friend Ronald Goldman. The trial was held in Los Angeles from January to October 1995. It has been described as the most publicized criminal trial in American history and educated a generation of Americans on the potential of DNA evidence. The trial included 133 days of televised testimony (available on the CNN website; People v. Simpson, Cal. Sup. Ct., 1995). The defense team persuaded the jury that there was reasonable doubt about the DNA evidence and cited mishandling of evidence by the Los Angeles police and sloppy internal laboratory procedures that contaminated the evidence. In 1997, a civil suit was brought against O.J. Simpson by the family of Ronald Goldman (Goldman v. Simpson, Cal. Sup. Ct.). The outcome was a $35 million wrongful death judgment against Simpson. Orchid Cellmark was the independent laboratory that tested more than 100 pieces of bloodstained evidence (Orchid Cellmark). In a different matter, O.J. Simpson was arrested in 2007 in Las Vegas and charged with armed robbery and kidnapping. He was found guilty and sentenced to a minimum of 9 years in jail without parole. Issues such as the novelty, reliability, and validity of routine DNA profiling in criminal trials in the ­twenty-­first century are rarely debated. However, on occasion, other issues may be explored by defense counsel. Over a decade ago, this author was required to serve as an expert witness in a contested trial involving DNA evidence in Melbourne, Australia (Case 3; Taupin, 2001).



Chapter one:  History of forensic DNA profiling in criminal investigations



9



Case 3 DNA was found on a plastic bag, business card, and a cardboard box containing ecstasy tablets that linked the accused to drug trafficking from the state of Queensland to Melbourne. The defense objected to the presentation of the evidence as “hearsay,” as the author of the report had not done all of the testing, and thus requested a voir dire examination on the admissibility of the evidence. In fact, the routine analysis of DNA evidence can involve six or more scientists and it is rare for the reporting scientist to be involved in all stages of the process. After ­cross-­examination by the defense on many aspects of DNA covering details of extraction and analysis, the judge admitted the evidence and the case proceeded to trial with the proviso that the first and second “typers” (who initially examined the DNA profile electropherogram) were to give evidence. The jury ultimately delivered a verdict of guilty. Many more case studies illustrating other aspects of DNA evidence in criminal trials will be discussed throughout subsequent chapters in this text.



1.5 Changing the face of forensic science: The value of biological evidence These highly publicized cases using DNA profiling led to an explosion in the use of DNA technology in criminal cases. This author has personally seen the value of analyzing biological materials in violent criminal cases increase exponentially since the introduction of such testing in cases handled by government and private forensic laboratories. The author’s MA thesis (Taupin, 1994) written early in the history of the use of DNA testing in criminal trials, however, showed that DNA is relevant in only a small proportion of criminal proceedings because of a lack of DNA evidence or because a matter is not contested (consent in sexual offenses, for example). Much of the impact is behind the scenes, when DNA profiling is used to exclude suspects or aid plea bargaining. DNA profiling technology is constantly evolving and improving as science continues to advance. The PCR innovation, for example, drastically changed the value of small and often degraded crime samples. Materials that cannot be seen by the naked eye, such as tiny bloodstains and sloughed skin cells, can now be revealed through DNA profiling. Innovative technology now allows DNA results to be obtained from



10



Introduction to forensic DNA evidence for criminal justice professionals



a single cell (Findlay et al., 1997). Note, however, that normal casework involves optimal requirements and specific limitations that will be discussed in Chapter 5 covering low levels of DNA. As of March 2013, the Innocence Project in New York has helped exonerate more than 300 people (many of whom were on death row) through DNA profiling. The leading contributors to these wrongful convictions were eyewitness identification problems and misleading forensic testimony. Kirk Bloodsworth was the first person sentenced to death whose conviction was overturned through DNA testing (Innocence Project; Case 4). Case 4 Bloodsworth was convicted in 1985 of the rape and brutal killing of a 9-­year-­old girl in Maryland in 1984. Although five eyewitnesses placed him with the girl, he continued to maintain his innocence. He read about the Pitchfork case in England while he was in prison and pushed to have the evidence in his case subjected to DNA testing. Initially, the evidence could not be located. Eventually the s­ emen-­stained underwear was found in the judge’s chambers and the semen DNA profiled. The DNA did not match Bloodsworth’s and he was released from prison in 1993 although not formally exonerated. Finally, in 2003, prisoner DNA samples were added to state and federal DNA databases and a match was obtained with Kimberley Shay Ruffner. A month after the 1984 murder for which Bloodsworth was convicted, Ruffner was sentenced to 45 years for an unrelated attempted rape/­murder and was incarcerated one floor below Bloodsworth. Ruffner pleaded guilty to the 1984 murder in 2004. Through DNA testing, the real perpetrators have been found in nearly 40% of these exonerations pursued by the Innocence Project. Some of this testing was performed after the original DNA tests were deemed faulty or the original forensic laboratory mishandled evidence. A perusal of the website profiles of the exonerations (Innocence Project) makes for informative reading for a criminal justice professional. This chapter has covered the continuing growth and strength of DNA evidence used in criminal cases. The following chapters will describe the techniques and also the strengths and limitations inherent in any DNA analysis performed at a crime scene, during a medical examination, or in a laboratory.



Chapter one:  History of forensic DNA profiling in criminal investigations



11



References Aronson, J.D. 2007. Genetic Witness. New Brunswick, NJ: Rutgers University Press. Avery, O.T., Macleod, C.T., and McCarty, M. 1944. Studies on the chemical nature of the substance inducing transformation of Pneumococcal types. Journal of Experimental Medicine, 79, 137–159. Findlay, I., Taylor, A., Quirke, P. et al. 1997. DNA fingerprinting from single cells. Nature, 389, 555–556. Franklin, R. and Gosling, R.G. 1953. Molecular configuration in sodium thionucleate. Nature, 171, 740–744. Gill, P., Jeffreys, A.J., and Werrett, D.J. 1985. Forensic application of DNA “fingerprints.” Nature, 318, 577–579. Gill, P. and Werrett, D.J. 1987. Exclusion of a man charged with murder by DNA fingerprinting. Forensic Science International, 35, 145–148. Goldman v. Simpson, Cal. Sup. Ct., LA County, Case SC036340. Human Genome Project. http://www.ornl.gov/sci/techresources Innocence Project. http://www.innocenceproject.org Jeffreys, A.J., Wilson, V., and Thein, S.L. 1985. Individual specific “fingerprints” of human DNA. Nature, 316, 76–79. Jeffreys, A.J., Brookfield, J., and Semeonoff, R. 1986. Positive identification of an immigration test case using DNA fingerprints. Nature, 317, 818–819. Jeffreys, A.J., Allen, M., Hagelberg, E. et al. 1992. Identification of the skeletal remains of Josef Mengele by DNA analysis. Forensic Science International, 56, 65–76. Mendel, G. 1865. Experiments in plant hybridization. English translation by C.T. Druery and W. Bateson. Journal of the Royal Horticultural Society, 26, 1–32, 1901. National Research Council. 1992. DNA Technology in Forensic Science. Washington: National Academy Press. Nature. http://www.nature.com Orchid Cellmark. http://www.orchidcellmark.ca/site/casefiles People v. Castro, 545 NYS 2d 985 (Sup. Ct. 1989). People v. Simpson, Cal. Sup. Ct., LA County, 1995, Case BA097211. R. v. Doheny and Adams, 1 Criminal Appeals R 369, 1997. State v. Woodall, 385 SE 2d (W. Va. 1989). Taupin, J.M. 2001. DNA analysis and drug trafficking. Forensic Bulletin, November 18 (Australia). Taupin, J.M. 1994. Impact of DNA Profiling on the Criminal Justice System. MA thesis, Department of Criminology, University of Melbourne, Australia. http://www.dtl.unimelb.edu.au/dtl_publish/research/28/66762 Venneman, M. and Koppelkamm, A. 2010. mRNA profiling in forensic genetics I: possibilities and limitations. Forensic Science International, 203, 71–75. Wambaugh, J. 1989. The Blooding. New York: William Morrow. Watson, J.D. 1968. The Double Helix. London: Penguin Books. Watson, J.D. and Crick, F.H.C. 1953. A structure for deoxyribose nucleic acid. Nature, 171, 737–738. Watson, J.D. and Crick, F.H.C. 1953. Genetical implications of the structure of deoxyribose nucleic acid. Nature, 171, 964–967.



12



Introduction to forensic DNA evidence for criminal justice professionals



Werrett, D.J., Lygo, J.E., and Sutton, J.G. 1989. The introduction of DNA analysis into Home Office laboratories in England and Wales, Banbury Report 32. DNA Technology and Forensic Science, 233–240. Wilkins, M.H.F., Stokes, A.R. and Wilson, H.R. 1953. Molecular structure of deoxypentose nucleic acids. Nature, 171, 738–740.



chapter two



Strengths and limitations of DNA profiling evidence 2.1 Introduction: Power and caution DNA profiling has become a ­well-­k nown technique used in criminal and other legal cases due to the massive publicity generated by high profile cases, TV crime shows, and films. DNA profiling is now used by organizations such as the Innocence Project in New York to rectify past errors in criminal cases. More than 50% of convictions that project has overturned were attributed to faulty forensic evidence presented in the original trials (Innocence Project). However, DNA must be present on evidentiary items for exoneration to be achieved. Exhibits from very old cases may have been lost, destroyed, or decomposed. Furthermore, no biological evidence may have been obtained in the original trial. Every legal professional, forensic scientist, and indeed member of the public should know that DNA profiling is neither infallible nor unassailable. Many documented instances demonstrate that contamination, transcription, and other errors have been made in DNA profiling (as described in Chapter 8) simply because the technique is performed and interpreted by human beings. It is also possible that DNA of an accused found at a crime scene may have an innocent explanation and/­or the DNA may not be relevant to the crime. Forensic means pertaining to the courts of law. The concept of science is less well understood, even by some practicing forensic scientists. Science is a method of study used to understand and describe the physical universe around us. The discipline of science is defined by the notion of hypothesis testing. First a hypothesis or theory is proposed. Experiments are then performed to test the hypothesis. The results of the experiments will either support or refute the hypothesis. The scientific method provides the framework for the testing of hypotheses. Alternative hypotheses should always be provided (Taupin and Cwiklik, 2010). Science is a dynamic endeavor whereby new discoveries change the way we think about the world. This is demonstrated in forensic science when the discovery of a more discriminating technique is used to distinguish samples that previously could not be differentiated. DNA profiling 13



14



Introduction to forensic DNA evidence for criminal justice professionals



is the best known example of a distinguishing technique. DNA profiling for use in forensic cases was a significant scientific achievement in the 1980s. It is an extremely powerful tool because it demonstrates very high discriminating power on samples from different individuals. The high exclusionary factor should be the focus of our attention, as all forensic science starts from an exclusionary perspective. Unfortunately the high statistics associated with the probability of a match have diverted attention away from exclusion and toward inclusion and matching. DNA profiling is perceived to be more objective than other forensic methods. Furthermore, it is robust and can be performed on ancient, poor quality, or very small samples, unlike the traditional techniques. No doubt these are the reasons behind its very quick uptake by the forensic and legal community; its notoriety in the solution of some high profile unsolved crimes also contributed to its widespread use. Accurate individualization is the aim of all forensic identification science. The problem is that the various types of forensic identification sciences do not use the same scientific paradigms or report their conclusions in the same format. DNA evidence is probabilistic and quantitative, while fingerprint evidence is categorical (match or n ­ onmatch). Forensic identification procedures may lead to categorical elimination, but unless the number of potential sources is limited and known, no forensic identification procedure can lead to categorical identification. A murder case from Arizona in the United States illustrates the consequences of unreserved acceptance of forensic evidence and represents a stark contrast between forensic analysis that has a solid scientific foundation and questionable forensic evidence (Innocence Project). Case 1 exempli­fies the advantages of applying DNA profiling to old cases adjudicated when routine blood grouping was not discriminatory. Case 1 The body of a waitress was found on a December morning in 1991 in a bar where she worked. She had been fatally stabbed and investigators focused on bite marks on her body. Upon hearing that the victim told a friend that Ray Krone helped her close the bar the previous night, the detectives asked Krone to make an impression of his teeth. Due to an accident, Krone had a distinctive dental pattern; experts stated that it matched the bite marks on the victim’s body. Saliva on the victim’s body came from someone with the most common ABO blood group, that of Krone. During trials in 1992 and 1996, Krone was convicted of murder and kidnapping, mainly due to bite mark testimony of an



Chapter two:  Strengths and limitations of DNA profiling evidence



15



alleged expert. Krone was dubbed the “snaggle tooth killer” and sentenced to life in prison. Then, in 2002, as a result of an investigation by Krone’s lawyer, DNA testing was conducted on the saliva found on the clothing associated with the bite marks on the victim. DNA profiling showed that the saliva did not match Krone’s. A search of the Combined DNA Index System (CODIS) database matched the DNA to that of Kenneth Phillips, an inmate at an Arizona prison. Krone was released in 2002 and Phillips entered a guilty plea in 2006. The expert then noted that the bite marks were not uniquely Krone’s and were consistent with the dental pattern of Phillips. The advent of DNA profiling has no doubt improved the recognition of biological evidence by investigators and the judicial system and led to questioning some unfounded assumptions. Case 2, also from the United States, is an example (Pilkington, 2012). Case 2 On September 29, 2012, Damon Thibodeaux became the 300th person to be exonerated by DNA evidence in the United States. He had been on death row in Louisiana since 1997 for the rape and murder of his 14-­year-­old step cousin. He had given a confession but much of what he said did not fit the facts. No semen was found on the body of the victim. The police theorized that semen was originally present but maggots destroyed it. Clothing from both Thibodeaux and the victim was reanalyzed and no DNA transfer between accused and victim was found. In fact, no evidence of rape was found and the maggot theory was absurd. Unfortunately, a mystique surrounds the utility of DNA profiling. The general public perception is that a suspect must be guilty of an offense if DNA evidence is found. This perception can extend to the criminal justice system and even forensic scientists. Gill and Buckleton (2010) noted in the forensic literature that “this is highly dangerous thinking.”



2.2 Discrimination power of DNA profiling The hypothesis testing should be an inherent part of a forensic examination. Forensic science comes from an exclusionary view. What makes a



16



Introduction to forensic DNA evidence for criminal justice professionals



forensic test so valuable is its discriminating power, that is, its potential for excluding. This is one reason why DNA profiling evidence is so powerful. Its discriminating power may be in the order of billions. Prior to the introduction of DNA profiling, the biological tests used to analyze forensic samples had very low discriminating power. The comparison of evidence samples to reference samples is a distinguishing characteristic of forensic science. Previously, routine serology such as ABO blood grouping was used to attempt to differentiate evidence and reference samples. Their discrimination levels were poor. For example, blood group A is common to a third of the population. For this reason blood grouping and other methods were most often used as exclusionary tools. However, samples taken from the same ABO blood group will have different DNA profiles and may be differentiated by DNA testing as shown in Case 1 above. Typical DNA profiling techniques examine at least 10 different areas on a DNA molecule and this procedure contributes to the very high chance probabilities. Like the other biological markers traditionally used, the advantages of DNA profiling include its genetic basis, backup from statistics, and availability of databases. The technique can be used to conclusively exclude an individual as the donor of an unknown source of DNA, if the DNA profiles from the individual and an unknown source are different. If the DNA profiles are the same, the statistical probability of obtaining a profile from a random individual in the population is determined. Crime stains, however, do not always yield full profiles (they may be partial) or single profiles (they may be mixed from two or more individuals) or they may be insufficient. The resulting lower discrimination and/­or evidential value may be crucial in a criminal case and these issues are discussed in Chapter 5.



2.3 Genetic basis for DNA profiling DNA profiling has a solid scientific foundation based on the science of molecular biology and genetics. The study of genetics and DNA is now part of high school biology studies and many texts cover the subjects. DNA profiling for forensic purposes is more specific and references will be given for each aspect covered in a chapter. A cell is the smallest working unit of a living organism. It consists of a liquid called cytoplasm that contains instructions for chemical processes, and a nucleus surrounded by an external membrane. The DNA molecule in the nucleus contains the genetic instructions for the development



Chapter two:  Strengths and limitations of DNA profiling evidence



17



and functioning of a living organism. Nuclear DNA is the DNA found within the nucleus of a cell and two types of DNA profiling use this DNA. Autosomal STR (short tandem repeat) profiling is the most common DNA typing system used in criminal cases and exploits the differences found on different chromosomes (see below for explanation of STR). Y-STR profiling examines only the Y chromosome in the nucleus, is inherited paternally, and has a specific purpose (see Chapter 6). Mitochondrial DNA (mtDNA) profiling examines DNA in mitochondria, found in the cytoplasm of the cell. MtDNA is inherited maternally and mtDNA profiling also has a specific purpose (see Chapter 7). A gene is considered a unit of heredity in a living organism. The location of the gene on the chromosome is called the locus and the different versions of a gene are called alleles. All humans have the same genes at the same loci on the same chromosomes, but the alleles differ from individual to individual. This makes every person unique from a genetic point of view. A genotype is the set of alleles of a gene. A homozygous genotype means that two identical alleles exist at the same locus, that is, the same allele was inherited from both parents. A heterozygous genotype reveals two different alleles at the same locus – different alleles were inherited from the organism’s mother and father. Generally, a female has the X chromosome only (denoted as X,X) and the male has one X and one Y (denoted X,Y). A DNA profile is the combination of genotypes obtained for different loci. It is important to remember that multiple loci are examined in DNA profiling to reduce the possibility of a coincidental match of unrelated individuals. Specific areas on the DNA molecule that are known to vary widely among individuals are examined using specific technical kits. The areas vary in length among different people and are called short tandem repeats (STRs). These repeat units are also called microsatellites and vary from 2 to 5 base pairs, typically repeated 5 to 30 times (Tamakin and Jeffreys, 2005). The STRs used in DNA profiling kits cover four base pairs. Analyzing the number of repeat units at many loci provides a highly sensitive measure of individual identity. STR markers are typically chosen to avoid problems of linkages between markers (see Chapter 4 on statistics). The STR markers need to be the same for DNA profile comparison within jurisdictions and across countries; however currently there is no one uniform marker set used although a set of “core STR loci” is used in Europe and the USA (see Section 2.7). The Amelogenin marker is the common sexing marker that determines whether X or Y or both is present.



18



Introduction to forensic DNA evidence for criminal justice professionals



There are separate criteria for determining the requirements for STRs in criminal casework and for database intelligence. Expanding a set of core STR loci: (a) reduces the likelihood of adventitious matches in the database; (b) increases international compatibility to assist law enforcement data sharing; and (c) increases the discrimination power to assist missing person cases (Butler and Hill, 2012) and thus is database focused. Criminal casework considerations require a focus on sensitivity, due to crime DNA samples often being compromised in some way (low level, degraded, or otherwise sub-optimal).



2.4 Stability of DNA profiling Routine serology markers used in forensic science before the advent of DNA profiling, such as ABO grouping or typing of enzymes, were often not applicable because the biological samples were small or the enzymes degraded quickly. The polymerase chain reaction (PCR) was a boon for forensic science, as it enables the analysis of minute quantities of blood and semen and is effective for degraded samples such as those commonly encountered at crime scenes. PCR is essentially a molecular photocopier that can amplify very small samples and allow them to be detected and analyzed. As biological samples age, the chemicals they contain begin to break down or degrade. Thus DNA will degrade, albeit more slowly than routine serology markers. In a biological sample, the interaction of the DNA and its environment is the factor that determines the preservation of the DNA. Age is not the only factor. Samples from a few years old to many decades old may still be analyzed successfully. The degradation process occurs minimally if the samples are preserved (kept dry and free from bacterial attack), but samples degrade rapidly when exposed to heat, moisture, or sunlight. It is thus important to ensure the proper collection and storage of biological evidence. Biological evidence has been found in improper storage facilities such as police locker rooms and courtrooms. Efficient storage of DNA extracts is also important. Reduction in DNA recovery has been observed after refrigeration of liquid DNA extracts and exposure to multiple freeze–­ thaw processes. The stability of DNA over time has been instrumental in solving cold cases and acquitting the wrongly convicted. Evidence from crime exhibits has been recovered and DNA profiled many years after offenses were committed. Case 3 involving Robert Dewey who was cleared of rape and murder in May 2012 is an example (Innocence Project).



Chapter two:  Strengths and limitations of DNA profiling evidence



19



Case 3 Robert Dewey was sentenced to life in prison in Colorado in 1994 for the rape and murder of a young woman found in her bathtub, partially clothed and strangled with a dog leash. Samples of blood found on Dewey’s work shirt were said to be a mixture of biological materials, some of which may have come from the victim. The bloody shirt and s­emen-­stained blanket found on the victim’s couch were stored by the police in an environment that prevented degradation. The items were retested with other evidence in 2011. The DNA from the blood on the work shirt was found to be Dewey’s. DNA from semen found on the victim’s blanket linked (via the national CODIS database) to Douglas Thames, then in prison for the rape and murder of another woman in 1989. Partial DNA profiles from materials under the victim’s nails and on the leash used to strangle her also matched Thames’s DNA. It is the opinion of this author that the transfer of DNA onto the body of a victim, especially in a rape and murder case involving transfer of appreciable amounts of DNA in semen and saliva, is a crucial factor in the solution of many cold cases. The clothing and medical exhibits from victims are often stored indefinitely and DNA profiling can assist in the resolution of cases many years after the crimes were committed. Human hairs, particularly the shafts, are less subject to degradation than blood or semen due to their construction and composition and may remain intact for centuries (Taupin, 2004). Items from exhibits thought to have been destroyed may still remain in laboratories or police files. Hairs on a microscope slide from a robbery were found in police rooms in the case of Sedrick Courtney after he was in prison for 15 years. Mitochondrial DNA showed he could not have been the offender (see Chapter 7 for case discussion). It is always worthwhile to investigate further if information indicates that original police evidence such as clothing from a victim has been destroyed. Medical swabs or debris collected from the items may still remain in case files, exhibit rooms, or refrigerators. The stability of nuclear and mitochondrial DNA makes it possible to obtain a result of evidential value many years after the commission of a crime.



2.5 Persuasive statistics The ultimate power of DNA profiling is its statistical discrimination. It is not possible to profile every person in a country, so a statistical probability



20



Introduction to forensic DNA evidence for criminal justice professionals



is performed if a “match” is found. The probability statistic is based on the frequency of each short tandem repeat (STR) at nine or more areas, and involves multiplication that gives rise to the high values. These values are in contrast to the relatively low values obtained by traditional serology tests. Relatedness of the population, sampling effects, racial databases, and validation studies are necessary before a statistic can be accepted. A probability of the rarity of the DNA profile or likelihood ratio is produced. A likelihood ratio compares the prosecution hypothesis against the defense hypothesis in statistical terms. See Chapter 4 for a more detailed discussion. An accused person may indeed be innocent and query the findings. One enterprising individual from the Northern Territory of Australia did so when told of a scientist’s opinion that “the chance of a second person having the same DNA profile was about one in 50 million.” The individual wanted to know what the police were doing about finding the second person (personal communication, George Georgiou, Victoria barrister). Case 4 illustrates the persuasive power of DNA statistics. The miscarriage of justice involved Farah Jama in Victoria, Australia and contamination was a factor. Case 4 The Victorian Parliament tabled a report about the 2008 conviction of Farah Jama for the rape of an unconscious woman in a nightclub (Vincent, 2010). Jama’s conviction was overturned on appeal in December 2009; he was initially sentenced to six years in jail and served 15 months. The inquiry found that DNA evidence was the only link between Jama and the woman. The report found that the offense probably never occurred and that medical samples from an unrelated sexual incident with another woman involving Jama (in which no charges were filed) were taken by the same medical officer at the same location within 30 hours of taking samples from the alleged rape victim. Jama’s DNA allegedly matched via a database hit when the samples arrived at the laboratory in an “unknown offender” case. Most likely, contamination between the evidentiary samples occurred during the medical examination, although the exact mechanism could not be determined. The recommendations of the report included education of legal practitioners and members of the judiciary on the nature and appropriate use of DNA. Justice Vincent stated in the report that the DNA evidence was perceived to appear so powerful by all involved in the



Chapter two:  Strengths and limitations of DNA profiling evidence



21



case that none of the filters on which our criminal justice system depends to minimize the risk of a miscarriage of justice operated effectively until weeks before Jama’s appeal. Vincent noted that no one appeared to be aware of the dangers of relying on statistical probabilities in the determination of guilt. A forensic scientist concluded that it was 800 billion times more likely that the DNA originated from Jama than from another Caucasian in the Australian population selected at random. Since the world population is only about 6 billion people, the 800 billion statistic appeared definitive. Jama subsequently received monetary compensation from the government.



2.6 Relatives Statistical analyses for DNA profiles are generally quoted for a population of unrelated individuals. When considering related individuals in a particular case, however, the statistics must be reviewed. For example, the offspring of a parent inherited one allele identical by descent, at each area profiled on a DNA molecule. The likelihood of observing a DNA profile in a related individual is thus higher than in the general population. Identical twins are particularly problematic, as they develop from the same egg and thus have the same DNA, as illustrated by Case 5. Case 5 In January 2009, jewelry valued at $6.8 million was audaciously snatched from a luxury department store in Berlin (Himmelreich, 2009). Three masked and gloved thieves were caught on a surveillance camera as they slid down ropes hung from the store’s skylights, thus outsmarting a sophisticated security system. DNA was found on a latex glove left at the scene, run through the national DNA database, and yielded matches with two people—identical twins. They were charged but released before trial because the court determined that at least one of the brothers was responsible but could not determine which one. See Chapters 4 and 5 for more detailed discussions of statistics.



22



Introduction to forensic DNA evidence for criminal justice professionals



2.7 DNA databases The United Kingdom became the first country in the world to launch a national DNA database in April 1995. New Zealand followed closely in 1996. The FBI implemented the Combined DNA Index System (CODIS) in the United States in 1998. In 2000, the CrimTrac National Criminal Investigation Database (NCIDD) was launched in Australia. However, not until 2009 did all jurisdictions in Australia use a single database. Today, national and regional DNA databases are in use through statute in many countries including those cited above. These databases have provided many cold hits for unknown samples left at crime scenes. Thus, in addition to comparisons with reference samples, biological evidence samples may also be compared to DNA databases to find potential matches with convicted offenders. To enable comparisons of DNA profiles both within and between jurisdictions, the same sets of loci must be analyzed. In 1997, the FBI selected a core set of 13 STR loci for use within the United States to upload DNA profiles to the national DNA database. Only 8 of these loci overlap with STR data gathered in the UK and most other European nations (Butler and Hill, 2012). Currently there is a 10 locus system used in the United Kingdom (AmpFLSTR SGM Plus™), 16 locus systems AmpFLSTR Identifiler™ and PowerPlex 16™ in the United States, and a 9 locus system AmpFLSTR Profiler Plus™ in Australia. Efforts are in progress to unify and expand the number of core STR loci for not only comparison across jurisdictions but to improve difficult database searches such as familial DNA where even 16 loci may not be sufficient (see Chapter 7). Profiling kits are being developed in response to a recommendation to expand the CODIS core loci from 13 STRs to 20 required loci and 3 optional loci (Hares, 2012). DNA databases generally comprise two groups of information: (1) DNA profiles of convicted offenders and volunteer samples, and (2) DNA profiles from criminal investigations (crime scene samples) stored in electronic format. For cases with no suspects, the crime scene DNA profile is loaded into the database to match against other profiles in specific categories: (prisoner, suspect, crime scene, limited purpose volunteer, unlimited purpose volunteer, missing person, and unknown deceased). Exclusions can help exonerate suspects in an investigation. A link provides the police with intelligence to assist in their investigations. The probability of identifying a suspect when a crime scene profile is checked on the U.K. database is greater than 40% (NDNA, 2009). The U.K. database contains about 5 million samples (about 10% of the population). This enormous number is a result of the relatively liberal criteria for entry into the U.K. database compared to other national databases.



Chapter two:  Strengths and limitations of DNA profiling evidence



23



Database matches are mainly through criminal offenses. However, a DNA database match or link does not represent an arrest or conviction; it merely indicates that a person may have been at a crime scene. Cases 1 and 3 from the United States, described above, show the outstanding utility of DNA databases in identifying the real perpetrators of violent crimes. Case 4 from Australia illustrates that matches may not be genuine. There has been an unacceptable rate of adventitious matches in database searches due to an increasing number of partial profiles in databases (Schneider, 2009). This is why there have been recommendations to incorporate “mini STRs” as part of the core loci in testing kits so that degraded or otherwise compromised samples may be more fully analyzed (see Chapter 5 for explanation of partial DNA profiles).



2.8 DNA ­intelligence-­led policing DNA is now a key component in ­intelligence-­led policing through: • Mass screening of suspects (such as in Case 1, Chapter 1*) • Targeting certain crimes in specific areas and collecting DNA evidence in an effort to catch repeat offenders; this is known as a targeted operation • National and international DNA databases • Familial matching Mass screening or DNA “dragnets” triggered privacy concerns in the United States and are not considered as productive as such operations conducted in Europe (Butler, 2012). There are three different types of DNA database comparisons. The first is the most common in criminal cases and is a comparison of the crime scene profile to the suspect’s profile, and would be accepted as the random match probability (say “p”). Database “trawls” are the second type, where the crime scene profile is searched for a match against “n”’ number of profiles in the database. There would be approximately “np” matches where no one in the database is the source and there is no relatedness in the database. The third, which is markedly different and is not used in criminal investigations, is the “trawl” through all possible pairs in a database so that every profile in the database is compared with every other profile and will result in greater numbers of matches (Kaye, 2009). The confusion in understanding the third type of database comparison is *



The Pitchfork case was the subject of Joseph Wambaugh’s 1989 book titled The Blooding.



24



Introduction to forensic DNA evidence for criminal justice professionals



a mathematical one based on “the birthday problem,” which is discussed in Section 4.3.4. The concept of familial matching is also controversial. A DNA profile from a crime scene may be searched against the DNA database of a country but only partial components of the profile may be used. This is because individuals in databases may have close relatives who have similar DNA profiles that potentially match DNA found at crime scenes. Chapter 7 discusses this topic.



2.9 Mass disasters Catastrophic events involving multiple victims and deaths are termed mass disasters. DNA profiling has proven invaluable in identifying victims of mass disasters such as the Boxing Day tsunami in 2004, one of the deadliest natural disasters in history. The epicenter of this earthquake in the Indian Ocean was off the coast of Sumatra in Indonesia, and affected most land masses bordering the Indian Ocean including Indonesia and Thailand. Countries such as Australia provided support that included DNA profiling identification of the victims. Smaller scale disasters such as air ballooning accidents may require DNA profiling for identification. Remains of the victims may be comingled, and identification in such instance poses particular problems. If bodies are not recovered quickly they degrade, especially in hot and humid climates. Bones or teeth may need to be considered for DNA analysis. If there is too much degradation, mitochondrial analysis and other techniques for analyzing “ancient” DNA may be required. A combination of DNA profiling techniques can be used to achieve positive identification, for example, of the last Romanov tsar, Nicholas II, and his family (see Chapter 7 for a discussion of this fascinating quest for identification).



2.10 DNA evidence in context The criminal justice professional should consider DNA results in the context of the case and ask several questions. How relevant is the DNA evidence? What is the quality of the other evidence? Is the DNA result the only evidence of substance? How many DNA results are there? Chapter 9 provides a list of questions to consider in determining whether to use DNA evidence. Sometimes no DNA results are generated. How do we interpret this? One possibility is that no DNA was transferred or deposited. This does not mean that the offender was not at the scene of the crime. The offender



Chapter two:  Strengths and limitations of DNA profiling evidence



25



may not have deposited or transferred detectable DNA. Principles of trace evidence transfer apply (see Section 2.11). Is DNA always paramount or relevant? Examining an assault case would require a search for blood on a suspect’s clothing, because blood evidence denotes injury to a victim. An examiner should also consider other evidence types according to observations and hypotheses. A DNA profile of blood on the suspect’s clothing that matches the DNA of the victim may support the presence of the suspect at the scene. However, if a suspect says he merely helped the victim after an attack by another person, the blood staining on the clothing of the suspect establishes his presence at the scene but is irrelevant in addressing the assault. The bloodstain patterns on the clothing may help disprove or support the hypothesis of the assault and the question of whether the suspect assisted or attacked the victim. Similarly, the presence of DNA from semen on a victim that matches the DNA profile of the accused is not the paramount factor in a case of rape fought on consent grounds. DNA can be obtained from handled objects, but the association of material that cannot be sourced to a body fluid such as blood or semen reduces the relevance of the DNA profile.



2.11 Time of deposition: Transfer and persistence of DNA DNA analysis has now advanced so that a profile can be produced from a single cell (Findley et. al, 1997). This means that we can analyze DNA that is not visible to the naked eye, and must consider issues such as the type of the material (blood, semen, skin cells), how the DNA may have been transferred, and how long it has been present at a crime scene or on an exhibit. The principle of trace evidence transfer is Locard’s Theorem, which is essentially “every contact leaves a trace.” This trace may or may not be detectable according to the type and amount of material in question and the method used. The principle is the basis for examination in many forensic textbooks to explain the rationale for the testing of exhibits (e.g., Taupin and Cwiklik, 2010). The potential for secondary transfer was first described in Melbourne, Australia. A study showed that plastic tubes held for fixed times by different consecutive users produced the DNA profile of the last holder and sometimes evidence of the DNA profile of the previous holder (van Oorshot and Jones, 1997). The authors warned of the potential for secondary transfer and contamination.



26



Introduction to forensic DNA evidence for criminal justice professionals



Primary transfer occurs when DNA is transferred from a person to an item. Secondary transfer is the transmission of that DNA to another item. Tertiary transfer occurs when the DNA on the second item is in turn transferred to a third. Many published studies have noted secondary transfers in quantities that can be detected from items that people simply touched to other items (Wickenheiser, 2002). DNA alone cannot be related to a specific action. DNA obtained from an object may have been deposited at separate times and places unrelated to the event in question if a body location cannot be determined. DNA from an unspecified cellular source reduces the relevance of biological evidence to a particular event and increases the uncertainty as to how the DNA may have been transferred to the item. Thus the relevance of findings may be difficult to assess (Gill and Buckleton, 2010). Chapter 3 provides more discussion on this topic. Case 6 comes from the author’s files and illustrates that the deposition of DNA does not necessarily tie an accused to a crime event. Case 6 An armed robbery trial was held in 2012 in the outback of New South Wales, Australia. The accused was an indigenous male. The trial was the second; the first trial ended in a hung jury. It was alleged that the accused threatened a shop attendant in a convenience store with a knife and a wrench, demanded cash, and left with money from the till. The attendant said the offender asked for and obtained a drink of water from a cooler inside the store. The offender then left two plastic disposable cups on the counter, one inside the other. Only one squashed cup was received by the laboratory. The top centimeter of the inner and outer rims of the cup was sampled. A full single source DNA profile matching that of the accused was obtained. The body origin of the DNA was not determined and was assumed by the laboratory to be saliva. The likelihood that the DNA profile would match an unrelated individual in the New South Wales population was estimated at 1 in 333 billion. However, statistics were not given for the indigenous population of the state (known to be more closely related than ­nonindigenous residents). Furthermore, only one cup was received by the laboratory, whereas two cups were left on the counter and examined by crime scene personnel. There was confusion about where the single cup originated. Empty plastic cups stacked on the water cooler inside the store appeared to be both used and unused.



Chapter two:  Strengths and limitations of DNA profiling evidence



27



The uncertainty as to which cup was used by the offender and at what time (not necessarily at the time of the offense), and a DNA match statistic not relating to the indigenous origin of the accused, led to a not guilty jury verdict. The number and type of sample impacts the evidential value of DNA in a case. A ­two-­way transfer with two separate DNA profiles (DNA in blood found on a suspect matches DNA of the victim and semen DNA on the victim matches DNA of the suspect, for example) is far more powerful than one DNA profile from an unspecified body origin.



2.12 Relevant evidence? It may not be possible to determine the time of deposition of DNA, so the DNA evidence may become less probative. How the DNA got on the exhibit and whether the DNA is relevant to the offense may also be paramount questions. The murder of Meredith Kercher in Italy (Case 7) was a high profile case covered by media around the world. The trial and appeal involved these factors in an interesting manner for criminal justice professionals (Hellmann, 2011). Amanda Knox, one of the three accused individuals, was from the United States. Raffaele Sollecito, her then boyfriend, was from Italy. The third individual was the Ivory ­Coast-­born Rudy Guede. Knox and Sollecito were freed on appeal in 2011 (Hanlon, 2011). The original verdict was criticized by one of the appeal court judges (Kington, 2011). In March 2013 Italy’s highest appeal court quashed the acquittals and ordered a fresh trial due to the manner in which the appellate court had been conducted (Davies, 2013). Case 7 A British exchange student, 21-year-old Meredith Kercher, was stabbed to death in Perugia, Italy in 2007. She was found in her bedroom in an apartment she shared with three other females including Amanda Knox. Kercher was found on the floor with stab wounds to her throat; she was sexually assaulted and some of her belongings had been stolen. Rudy Guede was convicted in 2008 of murdering and sexually assaulting Kercher and sentenced to 30 years, reduced to 16 years on appeal in December 2009. The evidence against Guede appeared uncontroversial, since DNA profiles matching his were found on Kercher’s body and clothing (Hellmann, 2011). The key DNA evidence



28



Introduction to forensic DNA evidence for criminal justice professionals against Knox and Sollecito was on bra clasps from the victim at the crime scene and on a knife found in a kitchen drawer at Sollecito’s flat. Knox was sentenced to 26 years and Sollecito to 25 years in December 2009. The conviction was quashed in 2011 through a successful appeal. The evidence of DNA defense experts Carla Vecchiotti and Stefano Conti was crucial to the appeal (­Conti-­Vecchiotti, 2011). The knife allegedly had traces of DNA matching Knox on the handle and DNA matching Kercher on the blade. The DNA alleged to have come from Knox was not disputed (it was found in her boyfriend’s flat) but the DNA profile alleged to have come from Kercher was very low level DNA (see Chapter  5 for a discussion on low level DNA and this case). There was no evidence that this low level profile was from blood. The suspects and victim knew each other and had access to each other’s apartments. It was not obvious why the knife was believed to be evidential. Questions were raised about handling and packaging of evidence. The bra of the victim had been cut and was found at her feet. The bra clasp was a fragment of material with a deformed clasp that had been removed from the rest of the bra and originally observed under Kercher’s body. The bra clasp displayed a clear major/­minor mixture profile (see Chapter 5 for mixture discussion). There was no dispute that the major DNA profile on the clasp came from Kercher (she had worn the bra). The minor DNA component was alleged to have come from Sollecito. Y ­ -­STR analysis (see Chapter  6 for discussion of the technique) showed a mixture from at least three males. The bra clasp was collected under a mat on the floor, more than a meter from its original position. It was also collected 46 days after the crime in a context highly suggestive of environmental contamination. The appeal panel consisted of six Italian citizens and two judges. The court ordered a review of the DNA evidence; the judges wrote that originally the scientific investigations occupied a preeminent position (Hellmann and Zanetti, 2011). The appeal court could not rule out contamination for the knife or the bra clasp and stated that low level DNA precautions did not appear to have been applied for the knife DNA. The court also noted an erroneous interpretation of both the autosomal DNA and ­Y-­STR profiles on the bra clasp.



Chapter two:  Strengths and limitations of DNA profiling evidence



29



This case is illustrative of the prosecution inferring the association of an activity such as stabbing or handling a bra clasp with a DNA profile that could not be sourced to a particular time or body fluid (similar to Case 6 above). Furthermore, collection procedures were issues in the decision. This case demonstrates the many factors that must be considered in the collection, handling, and interpretation of DNA evidence, particularly testing of low level or small amounts of DNA. When a DNA profile cannot be associated to a body fluid such as blood or semen, and collection practices are questionable (bra clasp collected 46 days after the crime), the criminal justice professional, whether acting for the prosecution or the defense, must evaluate the probative value of the DNA. Chapter 8 further discusses this case in relation to contamination and collection issues. Chapter 5 explains low level DNA techniques.



2.13 Relevant exhibits? The DNA evidence from the exhibits in a criminal case should accord with the hypotheses proposed, unless there is a plausible explanation. Sometimes the wrong exhibits are examined or the results may be insufficient. Case 8 illustrates the failure to consider all items submitted to a laboratory in the context of a case (Queensland Court of Criminal Appeal, 2001; Taupin and Cwiklik, 2010). Case 8 A 13-year-old girl was raped in Australia in 1999. She named Frank Button as the offender. Spermatozoa were obtained from vaginal swabs but no DNA profile could be obtained. Sheets and pillowcases from the girl’s bedding were also sent to the laboratory but not tested. Button was convicted and sentenced to seven years’ imprisonment. His appeal was heard on the grounds that no scientific evidence was presented at the trial. The laboratory then tested the bedding on insistence from the defense lawyers. The DNA profile from semen on the bedding did not match Button’s DNA profile. The laboratory then retested the vaginal swabs and obtained a DNA profile that also failed to match Button’s. The vaginal swab profile matched that of a convicted rapist and Button was acquitted. The criminal justice professional should be alert to the possibility of other exhibits that may have been collected by the police, may not have been tested forensically, and may be probative.



30



Introduction to forensic DNA evidence for criminal justice professionals



2.14 “CSI effect” and the notion of infallible forensic evidence The “CSI effect” is named after the eponymous television show, one of the world’s most popular TV shows that has spawned many sequels and imitators. These shows portray forensic scientists and crime scene investigators as clever and morally correct individuals who fight to clear the names of the innocent and put real criminals behind bars. The techniques portrayed on these shows also lead viewers to believe that they are viewing forensic science at work and that the science is infallible. However, the technology portrayed is intended for entertainment purposes instead of scientific accuracy. The worldwide legal fraternity believes the CSI effect has changed the way many trials are presented in that prosecutors are pressured to deliver more forensic evidence in court. Juries query the absence of forensic evidence and are likely to give more credence to prosecution cases that contain it. Defense lawyers say that the CSI shows make juries more unwilling to see that scientific findings can be compromised by human or technical errors. Although highly publicized, the reality of the CSI effect remains uncertain and we have no real evidence to conclude that jurors’ verdicts are distorted by it (­Goodman-­Delahunty and Verbrugge, 2010).



2.15 Relationships of lawyers and scientists Scientists and lawyers have a constrained relationship in the courtroom. Scientists may find it quite difficult to convey complex scientific principles in lay terms to lawyers and to juries. Defense lawyers may think they are opening “Pandora’s box” if they challenge the forensic evidence, for fear that it may appear even more infallible. There is a challenge in communicating successfully the true probative potential of DNA evidence to trial judges and jurors. Lay jurors are likely to need some guidance in making sense of evidence expressed in terms of probabilities. Chapter 9 provides guidelines for criminal justice professionals considering DNA evidence, along with a list of questions to ask themselves and their expert DNA witnesses.



References Butler, J. 2012. Advanced Topics in Forensic DNA Typing: Methodology. San Diego: Elsevier Academic Press. Butler, J. and Hill, C. 2012, Biology and genetics of new autosomal STR loci useful for forensic analysis, Forensic Science Review, 24, 15–26. Davies, L. 2013. Amanda Knox and Raffaele Sollecito face retrial over Meredith Kercher murder. The Guardian, March 26. http://www.guardian.co.uk/ world/2013/mar/26/amanda-knox-retrial-meredith-kercher-murder.



Chapter two:  Strengths and limitations of DNA profiling evidence



31



Findlay, I., Taylor, A., Quirke, P. et al. 1997. DNA fingerprinting from single cells. Nature, 389, 555–556. Gill, P. and Buckleton, J. 2010. A universal strategy to interpret DNA profiles that does not require a definition of low copy number. Forensic Science International: Genetics, 4, 221–227. ­Goodman-­Delahunty, J. and Verbrugge, H. 2010. Reality, fantasy, and the truth about CSI effects. InPsych, August. http://www.psychology.org.au/ publications/inpsych/2010/august/goodman Hanlon, M. 2011. As Amanda Knox walks free, now DNA evidence is on trial. Daily Mail Online, October 5. http://www.dailymail.co.uk/debate/article-​ 2044935/Amanda-Knox-freed-Now-DNA-evidence-trial-Kercher-murderacquittal.html Hares, DR. 2012. Expanding the CODIS core loci in the United States. Forensic Science International Genetics, 6, e52. Hellmann, P. 2011, The ­Hellmann-­Zanetti Report on the Acquittal of Amanda Knox and Raffaele Sollecito. http://hellmannreport.wordpress.com Himmelreich, C. 2009. Despite DNA evidence, twins charged in heist go free. Times Online. http://www.time.com/time/world/article/08599,1887111,00.html Innocence Project. http://www.innocenceproject.org Kaye, D. 2009. Trawling DNA databases for partial matches: What is the FBI afraid of? Cornell Journal of Law and Public Policy, 19, 145–171. Kington, T. 2011. Amanda Knox trial was flawed at every turn, says appeal judge. The Guardian, December 15. National DNA Database, United Kingdom. http://www.homeoffice.gov.uk/ science-­research/using-­science/dna-­database Pilkington, E. 2012. Louisiana death row inmate freed after 15 years—with a little help from DNA. The Guardian, December 7. Queen v. Frank Allan Button. 2001. Queensland Court of Criminal Appeal, QCA 133. Schneider. 2009. Expansion of the European Standard Set of DNA database loci— the current situation. Profiles in DNA, March. http//promega.com Tamaki, J. and Jeffreys, A.J. 2005. Human tandem repeat sequences in forensic DNA typing. Journal of Legal Medicine, 7, 244–250. Taupin, J.M. 2004. Forensic hair morphology comparison: A dying art or junk science? Science and Justice, 44, 95–100. Taupin, J.M. and Cwiklik, C. 2010. Scientific Protocols for the Forensic Examination of Clothing. Boca Raton, FL: CRC Press. van Oorshot, R.A. and Jones, M. 1997. DNA fingerprints from fingerprints. Nature, 387, 767. Vecchiotti, C. and Conti, S. 2011. ­Conti-­Vecchiotti Report. http://knoxdnareport. wordpress.com Vincent, Justice F. 2010. Inquiry into the Circumstances that Led to the Conviction of Mr.  Farah Abdulkadir Jama. Victorian Government Printer, Melbourne, Australia, 2010. Wambaugh, J. 1989. The Blooding. New York: William Morrow. Wickenheiser, R.A. 2002. Trace DNA: A review, discussion of theory, and application of the transfer of trace quantities of DNA through skin contact. Journal of Forensic Sciences, 47, 442–450.



chapter three



DNA profiling basics 3.1 What is DNA? Deoxyribonucleic acid (DNA) is a complex chemical found in the nuclei of all cells of the human body, except red blood cells. It is considered a genetic blueprint that is responsible for our chemical and physical characteristics. Half of an organism’s DNA is inherited from each parent—half from the mother’s egg that is fertilized and half from the father’s spermatozoa. Watson and Crick were awarded the Nobel Prize in 1962 for their discovery of the double helix structure of the DNA molecule. Their work was published in Nature (Watson and Crick, 1953). See Chapter 1 for a discussion of the history of DNA profiling. Each person’s DNA remains the same over his or her lifetime and the composition of the molecule remains the same throughout the body. This is a forensic advantage, because the DNA from a bloodstain at a crime scene can be compared with DNA from a reference saliva swab from a victim or suspect. Traditionally biological fluid typing, known as serology, was used as an investigative technique for solving violent crimes because biological materials are shed during violent acts. For example, blood is commonly found at homicide scenes and semen is found in rape cases. Today, DNA is more discriminating than traditional serology testing and it can be obtained from materials even when it cannot be seen.



3.2 Biological materials allowing DNA profiling Physiological fluids and biological materials are the most common types of physical evidence found in violent crime cases. The advent of DNA typing for individualization has increased its importance. Before the introduction of DNA profiling, the biological tests used to analyze forensic samples had very low discriminating power. Biological fluids such as blood, semen, and saliva were tested using ABO grouping and enzyme and protein tests that needed reasonable (visible) sample quantities and had low discrimination power. For example, blood group A is shared by one-third of the population. 33



34



Introduction to forensic DNA evidence for criminal justice professionals



Specific separate confirmatory tests are required to confirm the presence of blood, semen, or saliva. A DNA profile is specific to human material. However, to identify a body (somatic) location, a confirmatory test (such as the presence of spermatozoa to confirm semen) is required to associate a DNA profile to a biological material. The testing becomes even more complex if a sample contains a mixture of DNA profiles and potentially a mixture of biological fluids. A blood stain with a DNA profile mixture of at least two contributors must be considered in context. Could semen be present as well? Research is in progress to identify somatic origins from mixed samples and mixed DNA profiles. Blood and semen can be more readily associated with crimes due to their associations with violent or intimate contacts. However, trace DNA from an unspecified cellular source reduces the relevance of such biological evidence to a crime. Touch DNA is defined as arising from cellular materials in nucleated epithelial cells from the skin surfaces. The specific body location and time of deposition cannot be determined. Touch DNA may be obtained from handled objects but its presence is dependent on the amount of handling and the “sheddability” of the skin cells of the handler. Sheddability may be dependent on environmental conditions. DNA can be obtained from a wide variety of handled objects such as steering wheels, pens, cards, and bags, but the nature of these objects means that DNA found on them may have come from multiple contributors and thus produce mixed DNA profiles. The association of a specific body fluid and a DNA profile is not implicit (Taupin and Cwiklik, 2010; Peel and Gill, 2004). Currently DNA profiling can identify an individual from a sample of biological material but it does not reveal the body fluid or tissue source from which the profile originated. The determination of the type of body fluid is important both for evidential value and also to ensure the correct handling of samples. The amount of DNA borne per volume of sample material or exhibit varies according to the source. Solid human tissue and sperm samples of DNA contain very large amounts of DNA per unit volume. These samples should be considered as having the highest DNA potential when two or more types of ­DNA-­bearing cells are found on one item (Wickenheiser, 2002). Blood has the next highest DNA potential. Although blood is commonly found in violent crime cases, the ­DNA-­bearing white blood cells are outnumbered 400 to 1 by red blood cells. Saliva and nose and mouth secretions exhibit the third-highest DNA potential because of the small volume of body fluid conveying ­DNA-­bearing cells and small contact areas. Wearer DNA may be found on clothing but the quantity depends on the DNA deposited and the time elapsed. A garment may act like a reservoir of DNA if washed infrequently. Because of its transient nature, trace DNA



Chapter three:  DNA profiling basics



35



Human tissue



Sperm



Blood



Saliva



‘Wearer’ DNA



‘Touch’ DNA Figure 3.1  Relative DNA contents of biological materials found on crime exhibits.



has the lowest DNA potential. Figure  3.1 shows general order of DNA presence in human biological fluids.



3.2.1 Searching for DNA on exhibits The examination of biological evidence such as blood illustrates the steps in a forensic examination. The main objectives of biological evidence analysis are identification (or classification), individualization (DNA typing),



36



Introduction to forensic DNA evidence for criminal justice professionals



and reconstruction. The blood must first be located. This may be a difficult task if it is deposited in minute quantities or on dark colored surfaces. Presumptive and/­or confirmatory tests are then performed to identify stains or deposits such as blood that the examiner thinks may be of interest. Sampling will be an issue whenever an exhibit reveals multiple blood or other biological stains. How many stains should be sampled and how many tests should be performed? A DNA analysis result may be detrimental if all material is consumed in confirmatory tests. In any criminal case, the decision to sample or perform a particular test should be based on background information about the alleged crime and the scientific method used to test the hypotheses. The first step in identifying a body fluid is highly important since the nature of the fluid is itself very informative to the investigation. Furthermore, the destructive potential of a screening test must be considered when only a small amount of material is available. The ability to characterize an unknown stain at the scene of a crime without having to wait for results from a laboratory is another critical step in forensic body fluid analysis. Current tests for the identification of body fluids use chemoluminescence and the detection of specific proteins (Verkler and Lednev, 2009). Significant advances in laser technology and the development of novel light detectors have dramatically improved spectroscopic methods for molecular characterization over the past decade. Because gene expression patterns are tissue specific, a determination of the type of body fluid based on messenger RNA (mRNA) profiling may eventually be possible in routine case work. RNA can be isolated in suitable quality and quantity from blood, saliva, and vaginal secretions. A system using mRNA that can identify blood, saliva, semen, and menstrual blood in individual stains or in mixtures of body fluids has been developed (Fleming and Harbison, 2010). It should be remembered that blood stains may mask semen stains, and DNA from a sample may come from a number of sources. Mixtures from different body fluid sources and/­or individuals may appear on one exhibit or in a single stain (Taupin and Cwiklik, 2010). A mixed DNA profile from two individuals and two possible body sources may result. The analysis of semen on a complainant’s bed sheet may yield a mixed DNA profile. A major DNA component (or sperm fraction from differential lysis) of a stain likely to have come from semen may correspond to the DNA profile of a suspect. A minor DNA component likely to have come from epithelial cells may correspond to the DNA profile of a victim. It is not possible in these situations to determine whether the mixed stain occurred from biological fluids transmitted during sexual intercourse (Petricevic et al., 2006). Nor



Chapter three:  DNA profiling basics



37



is it possible to determine whether the DNAs from multiple individuals were deposited at the same time.



3.2.2 Blood Blood has traditionally been an excellent source of DNA (from the white blood cells). The quantity required for DNA analysis has decreased over the years as the technique has become more sensitive. Stains of 1 mm size or smaller can be analyzed as shown by Case 1. Case 1 A young boy named Damilola Taylor died in 2000 on a London housing estate as a result of a stab wound to his thigh that caused extensive blood loss at the scene (Rawley and Caddy, 2007). No forensic evidence was presented during the first trial of four boys in 2002. Two were found not guilty and charges were dropped against the other two. During a second police investigation, clothing belonging to two brothers, Danny and Ricky Preddie, were submitted for ­re-­examination at a different forensic laboratory from the one that initially examined more than 400 clothing items. A small drop of blood found on the heel of a white training shoe belonging to Danny Preddie was DNA profiled and found to match the profile of Damilola Taylor. A bloodstain was also found within the ribbing of a cuff of a sleeve of a black sweater belonging to Rickie Preddie. This stain also matched the DNA of Damilola Taylor. The discovery of the two bloodstains led to a prosecution of the two brothers and they were eventually found guilty of manslaughter in 2006. The Home Office review found that human rather than systemic failure led to the omission of examining relevant bloodstains in the first examination. The reviewers identified a conflict between the pursuit of excellence and the demand for urgent results. This case shows that sometimes only minimal evidence is found after a violent bloody attack. Although extensive blood loss may be obvious at a scene where a victim has bled to death, it may not be discovered for some time after the assault—long after the offender has fled. Blood from a stab wound may transfer very small quantities of blood to the clothing of an assailant. However, the age and quality of the stain will impact



38



Introduction to forensic DNA evidence for criminal justice professionals



the quality of a DNA profile. Old, degraded stains and those affected by molds and fungi may yield little or no DNA.



3.2.3 Semen and spermatozoa Semen obtained from a healthy male individual has a quantity of spermatozoa (sperm) that are rich in DNA. A specific extraction method is used to separate the sperm from other cellular materials such as vaginal cells obtained from medical swabs of the victim. Differential extraction of seminal fluid with spermatozoa aims to separate the seminal fraction (corresponding to the sperm of the male donor) from the cellular fraction that may correspond to the female from whom the swab was taken. Sometimes this extraction is incomplete or unsuccessful and female cellular material is found in the seminal (sperm) fraction and/­or lysed spermatozoa are found in the cellular (­nonsperm) fraction. This crossover of material from two separate donors can complicate interpretation. Sperm is destroyed quickly in the relatively hostile environment of a vagina. Many protocols recommend the taking of vaginal samples only if the ­postcoital interval is less than 72 hours or three days (­Mayntz-­Press et al., 2008). The literature notes that spermatozoa, although few in number, can sometimes persist in a vaginal canal longer than three days, but the survival rates are longer in the cervix. It has occasionally been found that spermatozoa survive more than a week in a deceased victim. The ability to obtain a DNA profile of a semen donor from a vagina using routine DNA profiling decreases rapidly after 24 to 36 hours following coitus and it is usually not possible to perform a profile after 48 hours (Saferstein, 2005). Sperm loss after intercourse is due to vaginal drainage, menstruation, and the normal sperm degradation that occurs over time. Loss can also occur during the multiple laboratory manipulations required for the differential technique used to separate the sperm from the n ­ onsperm DNA fractions. Case 2 illustrates the importance of the location of the semen in the context of a case. The DNA profile of the semen that matches the accused is not the only important factor. Case 2 The prosecution alleged that the accused had forcible vaginal sexual intercourse with the complainant. The accused stated that he ejaculated on the underpants of the complainant and that the act was consensual. Seminal stains were found on the front



Chapter three:  DNA profiling basics



39



of the complainant’s underpants, above the crotch line, and yielded a DNA profile that matched the DNA of the accused. The likelihood was estimated to be 1 in 10 billion that it was a random match with DNA from another individual. No semen or spermatozoa were detected in the medical swabs from the vagina of the complainant. The location of the semen stains did not correlate with the discharge of seminal fluid from the vagina, which would be expected on the gusset or crotch area of the underwear. Moreover, the absence of semen in the vaginal area did not accord with the proposition of sexual intercourse with ejaculation within the specified period of less than 24 hours. The accused was acquitted at trial. Source: Case study from author’s files. Internal (vaginal, anal, or oral) swabs from a sexual assault complainant are analyzed for spermatozoa using microscopic techniques. A medical officer often smears a medical swab onto a microscope slide so that the slide can be examined for spermatozoa, leaving the swab intact for subsequent examination. This may improve efficiency for the forensic scientist but creates potential for error and contamination because the evidence now consists of two separate entities. Major English laboratories create their own microscope slides from swabs to ensure a direct connection between sperm detected and the swab and also retain a representative sample for microscopy. It should be noted that it is possible to have penile penetration into a vagina without ejaculation of spermatozoa. A finger or other object may also penetrate without deposition of detectable DNA material. In all sexual offense cases, it is important that reference samples are obtained from both victim and suspect(s) for exclusionary purposes. Reference samples should also be obtained from any other male who may have had sexual activity with the victim during the few days before the alleged offense.



3.2.4 Saliva Nearly two decades ago, we learned that DNA can be extracted from saliva deposited on postage stamps (Hopkins et al., 1994). Obtaining DNA from saliva is now a routine process. The DNA obtained from saliva is not present in the salivary excretions themselves. It is present in the mouth (buccal) cells that are shed into the saliva. Thus, the DNA success rate on saliva is variable because it is not possible to predict the quantity of mouth cells in a saliva sample or stain.



40



Introduction to forensic DNA evidence for criminal justice professionals



On occasion, DNA may be recovered from drinking vessels, straws, food, cigarette and cigar ends, saliva stains on gags, targeted areas on headwear, envelope flaps, and licked stamps. The DNA from mouth cells is prone to degradation due to the high numbers of bacteria in the mouth. The DNA from saliva found in fizzy drink containers is particularly prone to degradation due to the acidic nature of carbonated drinks that may attack the DNA.



3.2.5 Hair roots Pulled hair samples that include the roots and/­or cellular material from the scalp and skin may be excellent sources of DNA. The roots of shed or fallen hairs that are often found on clothing or at crime scenes contain little cellular material. The hairs may provide mitochondrial DNA from the shafts. A number of successful solutions of cold cases were achieved by profiling DNA obtained from hairs in cases where little other evidence remained (see Chapter  7). Advances in technology allow examiners to obtain nuclear DNA from naturally shed hair roots and even from hair shafts. Thus hairs hold great evidentiary potential (Taupin, 2004).



3.2.6 Dandruff and skin Dandruff and the surface layers of skin may be suitable for nuclear DNA analysis. DNA may be obtained from any nucleated epithelial cells found in dandruff and the quantity has been estimated at 0.8 to 16.6 nanograms (ng) of DNA per dandruff particle (Herber and Herold, 1998). Medical conditions such as psoriasis and seborrhea dermatitis will also deposit skin in a similar manner to dandruff. Dandruff remains a problem among the normal healthy population. Fingernail scrapings may be of value when someone scratches another individual deeply enough to draw blood or remove skin.



3.2.7 Nasal secretions Used handkerchiefs can be valuable sources of relative large quantities of cellular DNA from the nose area. Also, an area of clothing such as a sleeve may been used by an accused or a victim to wipe his or her nose.



3.2.8 Vaginal secretions Vaginal fluid contains cells from the lining of the vagina. In a sexual offense case, vaginal fluid may be found on the outside of a condom or on



Chapter three:  DNA profiling basics



41



an item used to sexually assault a victim. Vaginal cells may also be present in a semen stain but cannot currently be differentiated from normal epithelial (skin) cells.



3.2.9 Sweat Sweat is a liquid secretion that contains no cellular material. However, certain areas of clothing such as the armpit of a shirt or the inner sole of a shoe may contain a mixture of sweat and skin cells sloughed from the body. It is believed that the sweat acts as a vector for the transfer of skin cells onto clothing. Areas of clothing may hold a reservoir of DNA and thus be valuable in identifying wearer DNA and determining the usual wearer of a garment.



3.2.10 Wearer DNA Wearer DNA is deposited on clothing when it is worn. This DNA is deposited through contact with the skin and consists of nucleated epithelial cells. The usual wearer of a garment should be detected as the major source of DNA on a garment, but minor DNA profiles of other individuals may also be detected if the wearer had close contact or lent the garment to another person. Nucleated cells from other body areas such as the eyes, nose, or mouth also yield successful DNA profiles. The hands may transmit nucleated cells to different parts of clothing. A cold case from Australia (Case 3) that baffled police for nearly 12 years illustrates the potential evidentiary value of wearer DNA belonging to an offender and the blood of a victim found on a pair of shoes—the blood caused the offender to discard them (Hall, Kennedy, 2011). Case 3 A heavily intoxicated 20-­year-­old man in Sydney, New South Wales became involved in a fight with a man with a goatee at a taxi stand in 1995. The victim was bashed and died from severe injuries in a nearby car park. Three unidentified men who came from a nearby nightclub were also involved in the fracas. The victim’s wallet was stolen from his back pocket and his running shoes were taken. Two days later, a lawyer in an office a few blocks from the murder scene looked out his office window and saw a pair of running shoes on an awning. The lawyer gave the shoes to the police and the DNA from the blood on the outside of the shoes matched the DNA of the victim in the bashing.



42



Introduction to forensic DNA evidence for criminal justice professionals The inside tongue of the shoes was sampled and another DNA profile was obtained from this wearer DNA. The profile did not match anyone on the database but it matched DNA from blood found on the inside of the victim’s back pocket. The police theorized the offender injured his knuckles in the altercation. He discarded his shoes because they were covered with blood and stole the victim’s shoes. In 2008, Darren Paul Smith was arrested by Queensland police for being drunk while riding a bicycle he had stolen after drinking in a pub. The larceny offense in Queensland required Smith to submit a sample for DNA profiling on the database. The sample placed on the national database matched the DNA profile from the wearer DNA on the tongue of the running shoes and from the back pocket of the victim’s trousers. The defense argued that the DNA in question came from transfer after the relevant exhibits were put into the same exhibit bag. The accused was found guilty by a jury and sentenced to a minimum of 18 years in prison.



3.2.11 Touch DNA The first demonstration that simply touching an object will leave sufficient amounts of DNA for a profile occurred in 1997 in a Melbourne laboratory (van Oorschot and Jones, 1997). The research arose from an unexpected DNA result in a criminal case. Since then, the amount of literature about investigating and utilizing touch DNA as evidence has risen exponentially. This author was involved in the investigation of the rape of an elderly woman in her own home (Case 4); touch DNA evidence led authorities to the offender (Taupin, 2009). Case 4 An unknown intruder armed with a knife raped an elderly woman in her home in northern England. She was dragged by the young male offender and subsequently found in the street outside, dressed only in her bra and sweater. She could not identify the offender beyond a general description. No semen was found on the house carpets, medical swabs of the victim, or the victim’s bloodstained sweater. Because she had been



Chapter three:  DNA profiling basics



43



dragged, there was potential for the deposits of touch DNA on the relatively rough surfaces of the lace parts of the bra cups and the areas were taped. The DNA analysis revealed a major DNA profile corresponding to the female victim and a complete minor male profile that was submitted to the national DNA database and matched one convicted offender. He had recently been released from prison and had just pawned three rings that belonged to the victim. He pleaded guilty before trial. It is recognized that some individuals may have more propensity to shed ­DNA-­containing cells than the rest of the population; these people are called “shedders” (Lowe et al., 2002). Knowledge of an individual’s shedding characteristics may be useful in compiling general background data in the interpretation of DNA trace evidence (Phipps and Petricevic, 2007). Because we cannot determine when touch dells are deposited, the degree of uncertainty about how the DNA may have been transferred to the object and the relevance of findings may be difficult to assess. In fact, the emphasis at trial may shift from “whose DNA is this?” to “how did this person’s DNA get here?” (Evett et al., 2002). The standard considerations applicable to all types of trace evidence (transfer, persistence, and recovery) are just as important for DNA evidence.



3.2.12 Urine and feces Urine and feces stains are not routinely examined because they do not usually contain enough cellular material sloughed from the lining of the urethra or anal canal to obtain a DNA profile. Both body materials are waste products. Only when sufficient cellular material may have been eliminated by a person (blood from hemorrhoids, for example) should a routine DNA analysis be considered. A scientist should be consulted and provided a specific scenario to determine whether obtaining DNA is feasible.



3.2.13 Emerging techniques Innovative and novel analyses are mentioned routinely in newspapers because they often capture the public imagination. Case 5 involved the use of maggots to identify the body they consumed (de Lourdes ­Chavez-­Briones et al., 2012). As a result, insects at crime scenes may now be investigated with more interest.



44



Introduction to forensic DNA evidence for criminal justice professionals



Case 5 In a wooded area, Mexican police found a body that was burned beyond recognition and identification was not possible even by DNA profiling. A woman had been abducted 10 weeks earlier and her ring was found nearby. The face and neck of the body were colonized by fly larva (maggots) that are frequently found and collected from corpses, especially those found outdoors. DNA typing was performed on the gastrointestinal contents of the maggots and compared to DNA obtained from the abducted woman’s father, with a probability of paternity of 99.685%. This was the first reported case of the use of human DNA from maggots to identify a victim in a criminal matter. Predicting externally visible characteristics such as hair and eye colors (phenotyping) is an emerging field in DNA profiling. The HIrisplex test that can predict both the eye and hair colors from DNA left at a scene was reported recently (Walsh et al., 2013). This test is expected to be useful when a perpetrator cannot be identified through DNA profiling.



3.3 Reference samples 3.3.1 Buccal scrapes The inside of the cheek is scraped four or five times with one or more plain sterile cotton or Dacron swabs to remove cells from the lining of the mouth. This type of sample is considered less invasive than the tra­ ditional reference blood sample and provides more than sufficient DNA if taken properly.



3.3.2 Blood Blood samples are often collected from deceased or surviving motor vehicle accident victims for medical and chemical tests and extra samples may be taken for DNA profiling.



3.3.3 Plucked hair samples Plucked hairs are collected at mortuaries as part of the autopsy protocol and can also be used for DNA profiling if the other body samples are degraded or otherwise unsuitable.



Chapter three:  DNA profiling basics



45



3.3.4 Personal belongings Missing person cases rely on items regularly used by the missing person such as toothbrushes and hair brushes. The DNA obtained from the belongings is then compared with DNA from relatives. Care should be taken in sampling belongings because they may have been used by someone other than the owner.



3.4 Current profiling technique: Short tandem repeats (STRs) The genetic markers currently typed in most forensic biology laboratories include autosomal short tandem repeats (STRs), mitochondrial DNA, and Y chromosome STRs (­Y-­STRs). This section will discuss nuclear DNA profiling using STRs (autosomal DNA profiling). The main steps in STR DNA profiling are • • • • • • •



Isolate the crime stain or other biological sample. Separate the DNA and clean the sample from the other material. Measure the quantity and quality of the DNA. Target the specific areas of interest within the DNA molecule. Produce multiple copies of the DNA pieces. Sort the DNA pieces according to size. Measure the sizes of the DNA pieces.



A major advantage of STR profiling is that many areas of a DNA molecule can be examined simultaneously in systems called multiplexes, thus reducing the amount of time required for a result. Forensic DNA profiling examines locations along a DNA molecule that are highly variable from one individual to another and still have no known functions. The original method of DNA profiling used by Alec Jeffreys was called restriction fragment length polymorphism (RFLP) and required relatively large amounts of DNA. The process was also complex and time consuming (see Chapter 1). The currently used polymerase chain reaction (PCR) technique was introduced in the 1990s and is useful for very small and degraded samples. PCR produces millions of copies of a particular area or sequence of DNA and ensures the production of sufficient material for analysis. It targets short repeated sequences or short tandem repeats (STRs) on a chromosome that are variable with specific sequence primers. The STRs are amplified many times so that they can be detected.



46



Introduction to forensic DNA evidence for criminal justice professionals



The DNA fragments that result from the process are then separated, detected, and analyzed. A major advantage of this type of STR profiling is that many areas of a DNA molecule can be examined simultaneously. It should be noted, however, that the potential for error also increases because of the many steps of the method and the human involvement at each step. The most common method of DNA profiling for forensic purposes uses a variation of short unit repeat loci (STRs) on chromosomes in the nucleus of the DNA molecule that are inherited from both parents. This is called autosomal STR profiling. These profiles are known as genotypes, as each STR is inherited from both the mother and the father. Humans have 22 pairs of autosomal chromosomes that are not involved in determining sex. The remaining pair consists of the ­sex-­determining X and Y chromosomes. The STRs (short unit repeat loci) in a DNA molecule are short unit lengths of DNA that are repeated end to end. A reasonable number of STR loci chosen (9 or more) can provide a high level of individualization in the population chosen for the sample. STR markers have become important tools for human identity testing and will continue to be used for many years because of their high degree of variability, ease of use in multiple amplifications, and implementation in national DNA databases. A core set of STR loci allows national and international sharing of criminal DNA profiles. STRs consist of regions of two to seven base pairs repeated in tandem. Individual variations involve the number of repeats and/­or the contents of the repeats. A variation in the content of the repeats occurs as a change in the base or as a deletion within a repeat unit. STRs used in forensics are either tetra (­four-­time) or penta (­five-­time) repeats. Most forensic laboratories use the ­four-­base pair repeat systems. STRs are highly abundant and well studied in the human genome, and their small size and the small size range of the alleles facilitate typing from highly degraded, small quantities of starting material. There are 9 core loci in the Australian DNA database system, 10 in the U.K. system, and 13 CODIS core loci in the United States database. Efforts are in progress to increase the number of loci examined in routine casework by using more sophisticated multiplexes such as the 16-loci Identifiler®Plus (Wang et al., 2012). The technique and the statistics used in autosomal STR testing are well developed and national and regional DNA databases are in use in many countries through statute. If no descriptive prefix precedes the DNA profiling term, it can be accepted that the nuclear STR method of DNA analysis described above has been used.



Chapter three:  DNA profiling basics



47



3.5 Reading tables of alleles A forensic DNA report may include a table of alleles that compares the crime scene sample DNA profiles with reference sample profiles. These tables most often also include samples from the accused and victim. The alleles are designated by a forensic scientist at a particular locus and correspond to the values detected at the locus for a specific profiling system. The SGM Plus™ system examines 11 loci including the amelogenin sex marker. Table 3.1 shows a table of alleles of three samples (one from the crime scene, one reference sample from the accused, and a victim reference sample). The crime scene sample also appears in the electropherograms of Figure 3.2 and Figure 3.3. The crime scene sample has all the designated alleles above the stochastic threshold (not low level, see Chapter 5). Across the top of the table are the names of the various loci examined. The alleles detected by the test at each locus are identified by numbers indicating short tandem repeats. An individual has two alleles at each locus, one inherited from each parent. In some cases, however, only one allele is detected. This is shown in Table  3.1 as 16,16 at locus vWA for the victim, and is interpreted as inheritance of the same allele (16) from each parent. Amelogenin (amel) is one of the loci analyzed and is used to determine the sex of the contributor. Males have X and Y chromosomes (X, Y in the allele table). Females only have X chromosomes (X, X in the allele table). An examination of the DNA profiles in Table  3.1 can determine whether the accused or the victim could or could not be the source of stain evidence. In this example, only the accused could have been the source. The table depicts a simple example, as the crime stain profile appears to involve a single source (no more than two alleles at each locus) from a male contributor. Both the accused and the victim are males (X, Y at the amelogenin locus). The forensic report associated with the above table would typically state that the victim was excluded from contributing to the crime scene sample and that the accused “matched” it and cannot be excluded. The report would also cite a statistic as to the probability of the match of the accused DNA reference profile and the crime scene DNA profile. The allele table is a summary of the results and does not show raw data or interpretation details. In order to analyze raw data such as the electropherogram (profile), the case file notes must be reviewed. The extraction and amplification dates, quantification values, and electropherograms of the crime samples and reference samples should be retained in the case file because the forensic DNA expert will have to explain this data to the criminal justice professionals involved in the case.



48



Locus Crime stain Complainant reference Accused reference



D3



vWA



D16



D2



AMEL



D8



D21



D18



D19



THO1



FGA



15,16 15,16 15,16



14,16 16,16 14,16



9,10 10,11 9,10



20,23 20,22 20,23



X,Y X,Y X,Y



12,15 11,12 12,15



28,31 29,29 28,31



12,15 12,15 12,15



14,15 14,14 14,15



7,9.3 7,9.3 7,9.3



24,26 24,25 24,26



Notes: The names of each locus are along the top row in their usual abbreviated form. The alleles of the crime stain are represented in the electropherograms of Figures 3.2 and 3.3. The full names of each locus are shown in the figures. The alleles are the numbers under each locus, except for the s­ ex-­determining amelogenin marker (AMEL). The accused is included as a contributor to the crime stain; the complainant is excluded as a contributor.



Introduction to forensic DNA evidence for criminal justice professionals



Table 3.1  Example of Allele Table Shown in Forensic DNA Report



110



100



120



130



140



vWA



150



160



15 16 100



AMEL



110



D8S1179



120



130



170



14 140



150



160



180



190



200



210



220



D16S539



230



240



16



170



250



260



270



280



D2S1338



290



300



9 10 180



D21S11



190



200



210



220



230



240



250



310



20 D18S51



260



270



280



290



320



330



340



350



23



300



310



320



330



340



350



300



310



320



330



340



350



160 120 80 40 0



X



Y



D19S433



200



100



110



12 13 120



130



140



28 150



TH01



160



170



180



190



200



31 FGA



210



12 220



230



240



250



260



270



280



Chapter three:  DNA profiling basics



D3S1358



240 200 160 120 80 40 0



15 290



160 120 80 40 0



14 15



7



9.3



24



26



Figure 3.2  Electropherogram (epg) of DNA profile. Three tones correspond to groups of loci recognized by fluorescent dyes. The loci are designated across the top of each color. D3 is the first locus and vWA is the second locus. The alleles are numbered peaks along the X (horizontal) axis, e.g., 15, 16 for D3. The X axis indicates the time the DNA fragments take to progress through the capillary. The Y (vertical) axis is measured in relative fluorescent units and denotes the amount of DNA present.



49



50



D3S1358



100



110



120



130



140



vWA



150



160



170



180



190



200



210



220



D16S539



230



240



250



260



270



D2S1338



280



290



300



310



320



330



340



350



2000 1000 0



15 16 AMEL



2400



100



110



D8S1179



120



130



14 140



150



160



16



170



20



9 10 180



D21S11



190



200



210



220



230



240



250



D18S51



260



270



280



290



23



300



310



320



330



340



350



300



310



320



330



340



350



1800 1200 600 0



X



Y



D19S433



100



110



12 13 120



130



140



28 150



TH01



160



170



180



190



200



31 FGA



210



12 220



230



240



250



260



270



280



15 290



1600 1200 800 400 0



14 15



7



9.3



24



26



Figure 3.3  Electropherogram from Figure 3.2 with a larger scale on the Y axis so that the tops of the peaks can be observed. Stutter peaks are present at all loci (but not designated) except THO1 and the amelogenin sex marker. The stutter peaks can be viewed more clearly in Figure 3.2 due to the smaller Y axis scale.



Introduction to forensic DNA evidence for criminal justice professionals



3000



Chapter three:  DNA profiling basics



51



The next section summarizes the basic steps in the DNA analysis of a sample that form the basis for the table of alleles or the statistic provided in a forensic report.



3.6 Obtaining DNA profiles The following steps are required for STR analysis of a biological sample found on an exhibit: • • • •



Extraction of DNA from biological sample Quantification of obtained DNA Amplification of STR loci by polymerase chain reaction (PCR) Separation and detection of amplification products by capillary electrophoresis



3.6.1 Controls It is essential that negative and positive controls and reagent blanks are processed through the analysis along with the sample in question to ensure that the quality systems work. The processing of these controls is required for every batch of DNA samples analyzed. A reagent blank in a DNA profiling examination tests for the possible contamination of the reagents and/­or supplies by an external DNA source during sample preparation. If the reagent blank exhibits one or more peaks above a certain threshold, it should be r­ e-­amplified. If the typing results remain after ­re-­amplification, all DNA samples associated with the reagent blank should be considered inconclusive and ­re-­extracted. If this is not possible due to the consumption of the DNA, the situation becomes a management issue. If the source of the contaminating DNA does not appear to be in the samples, the contamination should be noted in the report. If extraneous DNA is present in both the reagent blank and the associated sample, the sample should be reported as inconclusive. A positive control is used to determine the accuracy and consistency of the amplification and capillary electrophoresis processes. The positive control contains DNA from a known source with a known profile. If the positive control does not exhibit the appropriate results, the samples associated with the positive control should be considered inconclusive and ­re-­amplified. The negative control (amplification blank) contains all the reagents for the amplification mix but no DNA. The negative control tests for the contamination of samples during the setup of the amplification reactions. If the negative control exhibits unexplainable peaks above a certain threshold that are not eliminated after ­re-­injection, all samples associated with the negative control should be considered inconclusive and ­re-­amplified.



52



Introduction to forensic DNA evidence for criminal justice professionals



The negative control and reagent blank show whether contamination was present in the reagents or introduced through the testing process. The ­case-­reporting scientist should analyze the control results to ensure the data meet quality standards. A problem with controls should alert the criminal justice professional and it may not be resolved unless an expert witness peruses the raw scientific data.



3.6.2 Extraction The method of extraction of DNA from a sample depends on the nature of the sample. Epithelial cells from touch DNA require a simpler and quicker extraction to isolate the DNA; a more extensive method is required for spermatozoa and hair roots. Reference samples such as buccal swabs are simpler to extract than crime scene samples and have r­ ecipe-­based analysis protocols. Samples from crime scenes require scientific judgment as to the method of extraction according to body origin. Samples may be found on material that may inhibit a polymerase reaction (e.g., dyes on denim jeans) or be associated with a substance such as mold that may degrade the DNA. The DNA profile from such materials may reflect inhibition or degradation of the sample. Another extraction technique and a repeat analysis of the sample may be required to obtain an optimum profile. Evidence of degradation or inhibition of DNA in the sample may be observed in an electropherogram (see Section 3.6.5). The simplest steps in extraction begin with disrupting the cellular material to obtain the DNA and release other materials such as proteins. All the n ­ on-­DNA material is removed by adding detergents and proteases. The DNA material is often obtained as a pellet after a centrifugation (spinning down) process.



3.6.3 Quantification DNA purification methods cannot differentiate human DNA from other DNA, for example, from bacteria and fungi. If a sample is not pristine, a ­human-­specific DNA quantification system must be used. The purpose of the quantification step is to enable the addition of the optimum amount of DNA to the reaction to achieve amplification. This step uses a small part of the extracted DNA and compares it to a DNA standard of known quantification. Ideally, all samples will have the same amount of DNA added to the amplification mixture. Too much DNA will result in o ­ ff-­scale peaks, l­ ocus-­­to-­locus peak imbalance, and split peaks. Too little DNA will result in poor quality and low level profiles (see Chapter 5).



Chapter three:  DNA profiling basics



53



If no DNA is quantified, some laboratories stop an examination at this point. Other laboratories will proceed to the steps below in order to attempt a DNA profile. The quantification step is known to be less sensitive than the actual profiling step, so an attempt to achieve a result may be made if the evidence is considered crucial to the case, even if no DNA is quantified in this step. The absence of a quantifiable amount of DNA in a sample should be denoted in the case file because it indicates that low level DNA techniques may be required for further analysis and interpretation (see Chapter 5).



3.6.4 Amplification The two most important factors affecting amplification and success of nuclear DNA testing are the DNA quantity and degradation or inhibition of a sample. The amplification process is applicable only to the DNA of humans or higher primates (probably not an issue in criminal cases). If the amount of DNA extracted from forensic samples is too small to be detected by standard profiling, the amount must be increased. Amplification makes many copies of the DNA material at each locus. The technique for the amplification of the samples after the DNA is extracted and quantified is the polymerase chain reaction (PCR). It can be used on very small and degraded samples. It also targets the STRs on a chromosome that are variable with specific sequence primers. The STRs are amplified many times so that they can be detected. The DNA fragments that result from the process are then separated, detected, and analyzed. The PCR amplification procedure has three steps: • Denaturing: DNA strands of the double helix are separated by heating. • Annealing: By reducing the temperature, short synthetic DNA primers that flank the region to be amplified hybridize with the target DNA. • Extension: The temperature is changed again to an optimal level for the polymerase and new DNA is synthesized by polymerization. These steps constitute a cycle (generally 28 cycles but may be up to 34 for low copy numbers) and allow production of more than a million copies of the target DNA in a few hours. STR primers are ­human-­specific; if a profile is obtained, we can be assured that it is human. However, if no profile is obtained or sequenced, the DNA must be quantified to assure that it is human. The primers identify the relevant ­STR-­DNA segments and then amplify (replicate) these segments. Each primer is labeled with a fluorescent l­ight-­reactive dye to allow laser detection in the next step.



54



Introduction to forensic DNA evidence for criminal justice professionals



3.6.5 Separation and detection The fragments of DNA produced in the steps above are separated by capillary electrophoresis in a genetic analyzer. The fragments are forced by an electrical field through a narrow capillary tube in which the larger fragments move more slowly than smaller fragments. Under laser light, the colored dyes produce fluorescent light that signals the presence of DNA. A ­computer-­operated camera detects the light as the fragments reach the end of the capillary. Based on the color of the light and the time of travel through the tube, a series of computer programs determines which alleles are present at each locus. Dedicated software is used to interpret the c­ omputer-­generated data. The intensity and position of each light emission, displayed as a peak on an electropherogram (epg), is compared against standardized measures of known size and amount that constitute a sample known as a ladder that essentially serves as a reference. Peak heights are measured in relative fluorescent units (RFUs). The initial data produced by the fluorescent detection instrument is processed through software such as Genescan®. The raw data are then compared to the sizing ladder and peaks (alleles) designated by software such as Genotyper®. Other analysts (generally two) independently confirm the presence of the alleles and note other issues such as technical artifacts. The ­case-­reporting scientist is responsible for reviewing the software data and drawing conclusions about the data based on the quality of the profile and the controls.



3.6.6 Reading electropherograms The electropherogram (for example, Figure 3.2) can be viewed as a type of graph with an X axis (horizontal scale) and a Y axis (vertical scale). The positions of the peaks on the graph (distance to left or right) indicate how long it took a specific allele to pass through the capillary tube, and this indicates the length of the underlying DNA fragment. The numbers under each peak are ­computer-­generated labels that indicate which allele each peak represents and how high each peak is relative to the baseline. The smaller fragments are located toward the left side of the graph and the fragment sizes increase in length toward the right side. This is because it takes longer for larger fragments to migrate along the capillary tube than it does for the smaller fragments. The X axis is measured in time. The Y axis measures the intensity of the signal and thus the amount of DNA. As noted above, the RFU is the unit of measurement of the peak heights. It may be necessary to alter the ­printout scale if the peaks originally appear off scale on the Y axis. The heights of the peaks must be



Chapter three:  DNA profiling basics



55



determined (see Figure  3.3). Magnified versions such as Figure  3.2 are useful to determine the morphologies of smaller peaks that may not be observed readily in reduced versions such as Figure 3.3. Peaks representing alleles from the same person are expected to have roughly the same heights throughout a sample. This is certainly true for reference samples. However, in crime scene samples, degradation and inhibition may alter the balance of peak heights (see Section 3.6.7). Furthermore, mixtures of DNA from two or more contributors may also alter the proper balance (see Chapter 5). Accredited forensic laboratories will have their own validated DNA systems. The literature contains numerous ­peer-­reviewed articles, books, and reports on DNA analysis from extraction to electropherogram interpretation that are suitable as references for forensic examiners (Buckleton et al., 2004; Butler, 2005), although they are highly complex.



3.6.7 Artifacts and other technical issues Artifacts are peaks or other abnormalities in an electropherogram. Technical artifacts are observed often and have been documented extensively in the literature. Laboratories are required, as part of their quality systems, to use protocols to distinguish artifacts from real DNA peaks. An independent expert may be required for a second opinion if a profile of interest exhibits many artifacts. The presence of numerous dye blobs, spikes in the electropherogram, split peaks, and shoulders on peaks may indicate a poor quality profile resulting from poor or sloppy analysis. Artifacts such as stutter, however, can be expected. The most common stutter peaks are four base pairs smaller than the primary peak or associated allele. They result from a slippage of the strand during the amplification process, and are one repeat unit smaller than the designated allele on the electropherogram. Occasionally a forward stutter peak (four base pairs greater than its associated allele) will appear. Stutter peaks are evaluated by examining the ratio of stutter peak height to the height of the appropriate adjacent allele, expressed as a percentage. The height of stutter peaks can vary by locus but should not exceed 20% of each allele. Peaks in a greater stutter position may indicate a mixture of contributors to a profile. Figure 3.4 shows examples of stutter peaks immediately before each allele. Stutter peaks may also be observed in Figure 3.2. They are less discernible due to the size of the Y axis in Figure 3.3. Large stutter peaks, especially forward types, may indicate that too much DNA was analyzed for optimum results. Again, an expert should be consulted to determine the quality of the profile and decide whether the DNA extract should be ­re-­analyzed.



56



3000



290



300



310



320



330



340



2000



1000



0



20



23



Figure 3.4  Stutter in an electropherogram. The D2 locus has two designated alleles: 20 and 23. Each allele has one stutter peak four base pairs before it on the X axis that appear as a smaller peak.



Introduction to forensic DNA evidence for criminal justice professionals



D2S1338



Chapter three:  DNA profiling basics



57



­Pull-­up represents a failure of the analysis software to discriminate the different fluorescent dye colors labeled to the primers. ­Pull-­up can be readily seen on an electropherogram as the peaks are in the same X position as the designated allele from a different fluorescent dye. Again, ­pull-­up is most often observed when too much DNA is loaded onto a capillary tube. ­Pull-­up can be observed in Figure  3.2 and corresponds to allele peaks at the X axis (locus D16) and at FGA. The ­pull-­up in this electro­pherogram is at a low level and is not an issue. Spikes in an electropherogram are caused by fluctuation in voltage or air bubbles in the capillary tube. They do not look like allele peaks and should be readily visible. They are also usually observed in two colors. Dye blobs are usually broader than real peaks and are thought to appear when the fluorescent dye becomes detached from the DNA fragments. An ­off-­ladder allele may be marked OL on an electropherogram and may represent an unusual variant that does not accord with the ladder reference alleles. Alternatively, it could indicate a technical artifact. An interesting case in which an o ­ ff-­ladder allele was presented to numerous analysts who made different interpretations (Thompson, 2009) is presented in Chapter 5. Stutter and some other artifacts may complicate the interpretation of mixture DNA profiles by masking real peaks. This is also discussed in Chapter 5. Degradation of a DNA sample may often be observed on an electropherogram. Little or no degradation occurs if crime scene samples are well preserved and isolated from unfavorable conditions such as heat and moisture. The longer fragments of DNA are more likely to be affected first, and the consequence is that amplification may be partial or fail completely compared to results with shorter fragments of DNA. This may produce a “ski slope” effect in an electropherogram: the peaks toward the right side of the diagram are noticeably smaller in height than the ones toward the left-hand side. Sometimes the peak heights of larger fragments are too low to be discerned from the baseline and thus only a partial profile can be obtained. Degraded samples are particularly problematic in mixture samples because the two or more samples composing a mixture DNA profile may have different levels of degradation and thus lead to different interpretations by analysts. Figure 3.5 shows a degraded DNA profile; some of the loci show no discernible alleles. OL indicates an ­off-­ladder peak. The shorter fragments in Figure  3.5 (toward the left side) amplified better than the larger fragments to the right. The figure is a poor quality profile showing low peak heights and d ­ rop-­out. In the author’s opinion, the result should not be reported without further analysis. The sample may have been a ­t wo-­person mixture due to the presence of three alleles



58



110



100



120



130



140



vWA



150



160



170



180



190



200



210



220



D16S539



230



240



250



260



270



280



D2S1338



290



300



310



320



330



340



350



360



290



300



310



320



330



340



350



360



350



360



180 120 60 0



14 100



AMEL



110



D8S1179



120



18 130



16 17 18 140



150



160



170



180



12



D21S11



190



200



210



220



230



240



250



D18S51



260



270



280



400 200 0



X



Y



D19S433



400



100



110



14



11 120



130



140



150



28 TH01



160



170



180



190



200



30 FGA



210



12 220



230



240



250



260



270



280



OL 290



300



310



320



330



340



300 200 100 0



14 15



21 22



Figure 3.5  Electropherogram showing degradation and inhibition of DNA by reducing sizes of fragments from left to right in the diagram. An o ­ ff-­ladder allele is marked OL and a spike appears in the corresponding X position.



Introduction to forensic DNA evidence for criminal justice professionals



D3S1358



240



Chapter three:  DNA profiling basics



59



at locus vWA. The designated alleles at loci vWA, D16, D21, D18, and FGA are all below 200 RFU. No peaks are present for loci D2S and THO1. Inhibitors in the samples can affect the PCR amplification process. Body fluids left on soil, sand, wood, or vegetable matter can c­ o-­extract with human DNA and prevent or affect amplification. Other substances such as clothing dyes (denim jean dye is a notable example in this author’s experience) may contain polymerase inhibitors. Samples containing inhibitors often produce electropherograms similar to one from degraded DNA that shows the ski slope effect. It may not be possible to differentiate the cause of the small or absent peaks on the right side of the graph. One study explains the environmental and chemical degradation and PCR inhibition on single source samples and mixtures (McCord et al., 2011). There are alternatives to dealing with a compromised, poor quality DNA profile. The original DNA extract can be “cleaned up” via several methods (Butler, 2012). An extract colored with dye, for example, should indicate to an analyst that inhibition of the DNA extract is a real possibility.



3.7 Time required to obtain DNA profiles Stating a specific amount of time to produce a profile is difficult because it depends on the probative value of the evidence and importance and complexity of the case. I­ mpact-­based priority systems should be used to maximize the evidential value and allow the most crucial evidence to be examined first (Taupin and Cwiklik, 2010). A garment showing damage and multiple stains of blood and semen may require bloodstain pattern interpretation and damage analysis before testing for semen. Determining the sampling methods and correct stains for DNA profiling can be performed only after information is obtained from a physical examination of the evidence. The time for examination of the items and isolation of the pertinent DNA may vary from about 30 minutes to a few hours. The recovery of touch DNA is relatively simple and performed via tape lifting or swabbing an item. Recovery of semen from medical swabs or clothing is more time consuming because the presence of spermatozoa must be confirmed before analysis. This involves using a presumptive chemical screening test on presumed semen stains on a garment followed by microscopy to identify spermatozoa. Table 3.2 estimates the time required for each step of a DNA analysis. These times are based on samples designated urgent. Crime scene samples must always be analyzed with positive and negative controls to ensure the quality system works correctly. That is why a single sample cannot be analyzed in isolation to reduce the time waiting for results.



60



Introduction to forensic DNA evidence for criminal justice professionals Table 3.2  Analytical Times for DNA Stepsa



Step



Time



Extraction Quantification Amplification Separation and detection Analysis of fragments Interpretation



90 minutes for blood; overnight for semen and hairs 3 hours 3 hours 2 hours 2 hours for 2 scientists Variable, minimum of 1 hour



a



Approximate as of 2012.



Samples are usually processed in batches that contain samples from many cases. A batch of samples from multiple cases proceeds through the extraction, amplification, and other steps. Checking during the steps is vital to avoid transcription and other handling errors. Today, many forensic laboratories worldwide have large DNA analysis backlogs. This is the major factor that prevents prompt receipt of results in routine cases. Urgent cases, however, can be processed relatively quickly if an ­impact-­based priority system is followed. This author was the reporting scientist in a case in which DNA profiles (including wearer DNA) of three bloodstained items were obtained and interpreted in 1.5 days (information from author’s files). The case required overtime work by the examining scientist and the DNA analysts. Some forensic laboratories have backlogs of 6 to 12 months or more.



3.8 Designating peaks The peaks or alleles in an electropherogram or DNA profile are generally first designated in a forensic laboratory by automated software. Computer programs such as the GeneMapper® and Genotyper® are available to licensed laboratories. The programs utilize an allelic ladder that is essentially a sample that contains a sizing tool for reference against a crime DNA sample. After the peaks are designated by the computer program, the electropherogram is again interpreted by one or more laboratory scientists. The designated peaks are either affirmed or discarded according to the specific laboratory guidelines. Most often this is the point when the ­case-­reporting scientist compares the crime DNA profile with reference profiles. There are thresholds used in the interpretation of a DNA profile and the designation of peaks. The analytical threshold is a level above



Chapter three:  DNA profiling basics



61



which a peak may be determined as real and distinguishable from noise. Validation studies should be performed by the laboratory to determine the analytical threshold in use. Sometimes a threshold is given a uniform value of 50 or 100 RFU for every electropherogram. United Kingdom laboratories generally determine the analytical threshold for each electro­ pherogram based on the s­ ignal-­­to-­noise ratio. The stochastic threshold is another specification used in the interpretation of an electropherogram, particularly with low level profiles (discussed further in Chapter 5). Peaks below this threshold may exhibit ­drop-­out of one of the two alleles in a heterozygote. After a DNA profile is obtained, the case-managing scientist determines the next step. Often the next step is a comparison of a crime scene profile with one or more reference profiles. If the reference profile is excluded from contributing to the crime scene profile, then no statistic is generated. If the reference profile is included as contributing to the crime scene profile, a statistic should be generated. Chapter 4 discusses this topic. Although DNA profiling is considered more objective than other forensic studies, we can see that the discretion of a scientist still comes into play when interpreting DNA profiles. Chapter 8 discusses this subject in more detail. Examiner discretion applies most particularly to mixtures or low level DNA (see Chapter 5). Single source DNA profiles with more than adequate quantities should present little challenge. Chapter  5 also discusses an interesting case involving low level DNA presented to numerous analysts who provided varying interpretations (Thompson, 2009).



3.9 Case documentation and review Laboratory accreditation generally requires a technical review and an administrative review on every reported case. A technical review is essentially a peer review performed by a scientist not involved in the case. Independent conclusions are drawn from the data presented in the case file. Any discrepancies should be resolved by an independent third party (usually the manager of the unit). The administrative review is generally performed by the manager of the unit and ensures that all appropriate documentation is present in the case file, quality systems have been observed, and the final report makes sense and reflects the examinations performed. Figure  3.6 lists the contents of a generalized case file relating to a DNA report from a crime sample. If any of the listed documentation is missing, the legal professional should query its absence.



62



Introduction to forensic DNA evidence for criminal justice professionals



✓✓Forms for submitting exhibits to laboratory including case ID number ✓✓Chain of custody documentation for all exhibits from receipt to return ✓✓Requests for examinations by investigators ✓✓Details of reference DNA samples submitted for matter and reason for submission (identification of suspects, victims, known and accepted parties who may have deposited DNA, for example, a previous consensual partner in a sexual offense case) ✓✓Communications such as emails to and from laboratory biologists, investigators, and legal counsel ✓✓Scientific hypotheses formulated ✓✓Testing rationale including further testing if DNA results are not obtained for certain samples ✓✓Time frame proposed including projected court dates and urgent investigative deadlines ✓✓Examination notes relating to items of evidence (such as a pair of jeans) including photographs and diagrams ✓✓Presumptive test data (for example, acid phosphatase for determining semen) and results of searching for biological fluids ✓✓Confirmatory tests and results ✓✓Sampling details of exhibit for DNA analysis ✓✓Data sheets for each DNA sample including details of: • Extraction • Quantification (amount of DNA quantified in sample) • Amplification • Separation and detection • Accompanying controls • Electropherogram showing designated alleles and artifacts and thresholds used ✓✓Table of designated alleles for each sample ✓✓Determination of single source, mixture profile, partial profile, or low level for each sample ✓✓Determination of suitability of crime scene profile for comparison (indicate poor quality and/or insufficient alleles and possible need for reanalysis) ✓✓Comparison of crime scene samples with reference samples, bases for exclusion or inclusion, and reasons ✓✓If inclusion, note population database used for further comparisons and reasons ✓✓Statistical calculations • Software calculations • Manual calculations used to check software results (may be required in complex matters) • Control calculations (dummy data) to ensure software and manual calculations work as expected ✓✓Technical review notes ✓✓Administrative review notes ✓✓Final copy of scientific report



Figure 3.6 File documents that should be retained in cases involving forensic DNA testing. They will be required for forensic reports.



Chapter three:  DNA profiling basics



63



References Buckleton, J.S., Triggs, C.M., and Walsh, S.J., Eds. 2004. Forensic DNA Evidence Interpretation. Boca Raton, FL: CRC Press. Butler, J. 2012. Advanced Topics in Forensic DNA Typing: Methodology. San Diego: Elsevier Academic Press. Butler, J.M. 2005. Forensic DNA Typing: Biology, Technology, and Genetics of STR Markers, 2nd ed. Burlington, MA: Elsevier Academic Press. De Lourdes ­Chavez-­Briones, M., ­Hernandez-­Cortes, R., ­Diaz-­Torres, P. et al. 2012. Identification of human remains by DNA analysis of the gastrointestinal contents of fly larvae. Journal of Forensic Sciences, online September 12. Evett, I.W., Gill, P., Jackson, G.M. et al. 2002. Interpreting small quantities of DNA: The hierarchy of propositions and the use of Bayesian networks. Journal of Forensic Sciences, 47, 520–530. Fleming, R.I. and Harbison, S. 2010. The development of a mRNA multiplex ­RT-­PCR assay for the definitive identification of body fluids. Forensic Science International: Genetics, 4, 244–256. Gill, P. and Buckleton, J.S. 2010. A universal strategy to interpret DNA profiles that does not require a definition of low copy number. Forensic Science International: Genetics, 4, 221–227. Hall, L. 2011. Drunken cyclist’s DNA leads to cold case murder conviction. Sydney Morning Herald, September 17. Herber, B. and Herold, K. 1998. DNA typing of human dandruff. Journal of Forensic Sciences, 43, 648–656. Hopkins, B., Williams, N.J., Webb, M.B.T. et al. 1994. The use of minisatellite variant repeat polymerase chain reaction (­MV-­PCR) to determine the source of saliva on a used postage stamp. Journal of Forensic Sciences, 39, 526–531. Kennedy, L. 2011. If the shoe fits … killer nabbed 16 years later. Sydney Morning Herald, July 10. Lowe, A., Murray, C., Whitaker, J. et al. 2002. The propensity of individuals to deposit DNA and secondary transfer of low level DNA from individuals to inert surfaces. Forensic Science International, 129, 25–34. ­Mayntz-­Press, K.A., Sims, L.M., Hall, A. et al. 2008. ­Y-­STR profiling in extended interval (≥ 3 days) postcoital cervicovaginal samples. Journal of Forensic Sciences, 53, 342–348. McCord, B., Opel, K., Funes, M. et al. 2011. An investigation of the effect of DNA degradation and inhibition on PCR amplification of single source and mixed samples. National Institute of Justice. http://nij.gov/publications/digest/ issue6.htm Peel, C. and Gill, P. 2004. Attribution of DNA profiles to body fluid stains. International Congress Series, 1261, 53–55. Petricevic, S.F., Bright, J.A., and Cockerton, S.L. 2006. DNA profiling of trace DNA recovered from bedding. Forensic Science International, 159, 21–26. Rawley, A. and Caddy, B. 2007. Damilola Taylor: An Independent Review of Forensic Examination of Evidence by the Forensic Science Service. London: United Kingdom Home Office. Saferstein, R., Ed. 2005. Forensic Science Handbook, 2nd ed., Vol. 11. New York: Pearson Prentice Hall.



64



Introduction to forensic DNA evidence for criminal justice professionals



Taupin, J.M. 2004. Forensic hair morphology comparison: A dying art or junk science? Science and Justice, 44, 95–100. Taupin, J.M. 2009. Targeting areas of clothing for ‘touch’ DNA: Unrelated clothing damage analysis leads to the solution of an aggravated rape. European Academy Forensic Sciences Meeting, Glasgow, September. Taupin, J.M., and Cwiklik, C. 2010. Scientific Protocols for Forensic Examination of Clothing. Boca Raton, FL: CRC Press. Thompson, W.C. 2009. Painting the target around the matching profile: The Texas sharpshooter fallacy in forensic DNA interpretation. Law Probability Risk, 8, 257–276. van Oorschot, R.A.H. and Jones, M.K. 1997. DNA fingerprints from fingerprints. Nature, 387, 767. Virkler, K. and Lednev, I.K. 2009. Analysis of body fluids for forensic purposes: From laboratory testing to n ­ on-­destructive rapid confirmatory identification at a crime scene. Forensic Science International, 188, 1–17. Walsh, S., Liu, F., Wollstein, A. et al. 2013. The HIrisPlex system for simultaneous prediction of hair and eye colour from DNA. Forensic Science International: Genetics, 7, 98–115. Wang, D.Y., Chang, C.W., Lagace, R.E. et al. 2012. Developmental validation of the AmpFlstr® Identifiler®Plus PCR amplification kit: An established multiplex assay with improved performance. Journal of Forensic Sciences, 57, 453–465. Watson, J.D. and Crick, F.H.C. 1953. A structure for deoxyribose nucleic acid. Nature, 171, 737–738. Watson, J.D. and Crick, F.H.C. 1953. Genetical implications of the structure of deoxyribose nucleic acid. Nature, 171, 964–967. Wickenheiser, R.A. 2002. Trace DNA: A review, discussion of theory, and application of the transfer of trace quantities of DNA through skin contact. Journal of Forensic Sciences, 47, 442–450.



chapter four



Evidential value and statistics 4.1 Introduction The ultimate power of DNA profiling is in its statistical discrimination. It is not possible to perform a DNA profile of every person in a country, so a statistical probability is determined by a scientist if a match is found. This statistic is based on the frequency of each allele or STR at nine or more areas (depending on the profiling system used in the jurisdiction) and involves multiplication that gives rise to the high values often observed. However, factors such as relatedness of the population and sampling effects are most often used in the overall calculation. Statistics in general, as mentioned in Chapter 2, can cause difficulty in interpretation for investigators, lawyers, and others in the criminal justice system such as jury members. The advent of DNA profiling and its inherent reliance on statistics led various organizations to recognize this problem and produce guidelines for legal advocates. A United States judicial body developed a manual on scientific evidence (U.S. Federal Judicial Center, 2011). The manual was written with the needs of a legal audience in mind and covers a range of related topics including: data collection and presentation, comparisons, association and causation, and DNA evidence. The Royal Statistical Society in London also produced two practitioner guides for lawyers litigating DNA evidence (Aitken et al., 2010; ­Puch-­Solis et al., 2012). Aitkin’s guide focuses on statistical analysis; P ­ uch-­Solis et al. cover DNA evidence. The statistical concepts in DNA profiling evidence may be hard to grasp, particularly for those with little mathematical background. The author’s experience finds that the concepts are difficult to communicate without setting aside a considerable period of time for contemplation. The criminal justice professional should try to understand the basic concepts in DNA profiling but should also seek clarification regarding the statistical meanings of the evidence in his or her case from the forensic DNA expert. The next section summarizes the main statistical issues in DNA profiling evidence along with some pertinent case studies. More complex cases involving mixtures from two or more unknown individuals, low 65



66



Introduction to forensic DNA evidence for criminal justice professionals



level DNA samples, and partial DNA profiles are addressed in Chapter 5. ­Y-­STR profiling and mitochondrial DNA both require different statistical considerations. These are discussed in Chapters 6 and 7. Case 1 in Section 4.3.3 briefly covers an interesting case.



4.2 Interpreting DNA profiles DNA profiling is a comparative technique. A laboratory compares the DNA result of a crime sample with that obtained from a reference sample. If the two profiles are different, the donor of the reference sample cannot have shed the body fluid from which the crime sample was generated. If the DNA profiles are the same, the result is a match. One of the presumptions in determining a match between a crime DNA profile and a reference DNA profile is that the peaks (or alleles) are designated correctly. Three kinds of alleles appear in a crime stain profile (Gill et al., 2006):



1. Alleles that are unmistakable 2. Alleles that may be masked by artifacts 3. Alleles that have dropped out completely and cannot be detected



Points 2 and 3 are discussed in Chapters 5 and 8. The subjectivity of the analyst is certainly a factor in designating a peak (Dror and Hampikian, 2011). Ensuring that appropriate biochemical and genetic tests are performed will mean that the best result is obtained. It is better to analyze further or replicate stains on an item to try to obtain a better result than to perform statistical analyses on suboptimal results. If two profiles match, the person who provided the sample or someone who also has the same DNA profile can be the source of the evidential material. The significance of the match is determined by a statistical analysis. An ­in-­depth discussion of the statistical approaches is beyond the scope of this book but many detailed texts on this subject are available (Aitken et al., 2010; Balding, 2005; Evett and Weir, 1998; ­Puch-­Solis et al., 2012). The criminal justice professional is encouraged to be familiar with the basic principles and some of the terminology. Appendix A is a glossary of common terms.



4.3 Statistical approaches and obtaining final statistics 4.3.1 Random match probability and likelihood ratio The interpretation of DNA profiles when a match is found requires a determination of the probability that a second copy of the DNA profile



Chapter four:  Evidential value and statistics



67



will be present in a certain population. The forensic literature contains much debate about how this probability should be derived. The two methods in common use to report DNA profiles are the classical probability approach and the likelihood ratio approach (Buckleton, 2005). An appendix at the end of a laboratory report should provide information about the derivations and meanings of the statistics applied to evidence implicating the suspect. A laboratory analyzing evidence yielding full single-source DNA profiles will use one of two statistics: 1. Random match probability based on genotype frequency estimates 2. Likelihood ratio based on the primary hypothesis that the suspect is the source of the DNA profile versus the alternate hypothesis in which an unrelated individual from the general population is the DNA donor







Random match probability (RMP) is the chance of a random DNA profile match within a given population and is the reciprocal of the DNA profile frequency. A DNA profile frequency is estimated by determining the genotype frequency for each locus and then multiplying the frequency across all loci. Rare genotypes provide stronger evidence, and population databases sorted by race will yield somewhat different results. However, it is important to understand that the result is a representation of how rare a DNA profile is in a population. The profile probability approach presents the probability of the occurrence of an evidentiary DNA profile (E) under a stated hypothesis H0. This hypothesis may be as simple as saying that the DNA profile is from a person unrelated to the suspect. The probability is written formally as: Pr (E/­H0) where Pr is the abbreviation for probability and the conditioning bar is an abbreviation for given. Under the approximation that profiles from unrelated people are independent, this probability for a single stain is the frequency of occurrence of the profile in the population. An extension of the profile probability approach works with the probabilities of the evidence under two or more alternative hypotheses about the source(s) of the profile. This is called the likelihood ratio (LR). A typical analysis of a crime sample utilizes the prosecution hypothesis Hp (the accused is the source of the DNA) and the defense hypothesis Hd (the accused is not the source of the DNA). The LR is becoming the preferred approach worldwide. If the LR is greater than one, the evidence supports the first (prosecution) proposition;



68



Introduction to forensic DNA evidence for criminal justice professionals



if it is less than 1, it supports the second (defense) proposition. The LR can cope with other factors such as the uncertainty about the number of contributors in a mixed profile (see Chapter 5 for mixture interpretation). An overall composite likelihood ratio is obtained by multiplying the likelihood ratios for each locus on the assumption that the defendant’s profile matches the questioned profile for the prosecution hypothesis.



4.3.2 Calculating frequencies It is necessary to determine the genetic composition of the relevant population with allowances for ­coancestry, sampling, and blood relative issues. The frequency of genotypes among major populations in the relevant location may have been determined and published in forensic journals or may be calculated by a particular laboratory if it maintains a particular ethnic data set and its results have been published and validated. The size of a database for these calculations may be as small as 100 individuals and still be valid for making reliable projections about a genotype’s frequency in a larger population (Chakraborty, 1992). The adequacy of the sampling allowance method and the allele counts should always be assessed formally (Curran and Buckleton, 2011). The “product rule” is the simplest statistical calculation regarding DNA evidence and was developed from Mendel’s work (discussed in Chapter 1). If the particular population from which the DNA is postulated is large enough, it is assumed that any random effects can be ignored. The independence of the loci (DNA molecule areas from the profile), also known as the ­Hardy–­Weinberg equilibrium and linkage equilibrium, is assumed. Allelic frequencies from databases that are deemed to meet the ­Hardy–­Weinberg criteria of independence and random mating are used to calculate the genotypic frequencies of each STR locus result. These genotypic frequencies are then multiplied together to generate an estimated frequency of occurrence of the obtained DNA profile in the population to which the database corresponds. Likelihood ratio calculations incorporate factors such as the inbreeding coefficient of a particular ethnic database and sampling correction. The calculation is based on the Bayes theorem and conditional probability. The Bayesian approach is now the foremost alternative for forensic disciplines, in the literature if not in actual practice. Conditional probability can be stated simply: given that A occurs, what is the probability that B occurs? Stated another way, the probability of B is conditioned on the occurrence of event A. The Bayesian approach is based on at least three ideas:



Chapter four:  Evidential value and statistics



69







1. It is necessary to consider an alternative proposition in any evaluation of a probability. 2. Scientific interpretation is based on the probability of the evidence given the proposition. 3. The interpretation is also conditioned on the framework of circumstances. The Balding and Nichols sampling formula is a correction factor incorporated into the statistics for the likelihood ratio. It accounts for the fact that the frequency of the particular genotypes in a laboratory population database came from a sample of the population, not the entire pop­ ulation (sampling an entire population is not currently possible).



4.3.3 Comparison of probability of exclusion and LR methods The combined probability of inclusion (CPI) and the combined probability of exclusion (CPE) calcula­tions are used by some laboratories to indicate the statistical significance of results. CPI is the percentage of the population that can be included in a profile; CPE is the percent­age of the population that can be excluded. The CPI and CPE calculations are closely related: CPI is calculated by multiply­ing the probabilities of inclusion from each locus, and CPE is calculated by subtracting the value obtained from the CPI calculation from 1 (i.e., 1 − CPI). This terminology is most often observed in ­Y-­STR typing and mitochondrial DNA reports because likelihood ratios cannot be performed for these analyses (see Chapters 6 and 7). Case 1 (Ayturgrul v. The Queen, 2012) was an appeal to the High Court in Australia after a DNA statistic was alleged to be inadmissible. Case 1 The appellant was tried in the Supreme Court for murder and was found guilty. The deceased and the appellant had been in a relationship that ended more than two years before the victim was stabbed to death. The prosecution case at trial was circumstantial. Mitochondrial DNA testing (see Chapter 7) was performed on a hair found on the deceased’s thumbnail. This test showed that the accused “could have” been the donor of the hair and two statistics were submitted. It was expected that 1 in 1,600 people in the general population would share the mitochondrial DNA profile or that 99.9% would be excluded. It was alleged that the exclusion percentage was not permissible.



70



Introduction to forensic DNA evidence for criminal justice professionals The Court of Criminal Appeal dismissed the appeal and special leave was granted to appeal to the High Court that then held that the appellant did not demonstrate that the probative value was outweighed by the danger of unfair prejudice.



It should be noted that mitochondrial DNA typing and Y ­ -­STR profiling use different techniques from autosomal STR DNA profiling, and the derivation of the statistical significance is different. The techniques are less discriminatory than autosomal STR DNA profiling due to the method of inheritance of haplotypes—either from the maternal line (mitochondrial) or from the paternal line (­Y-­STR). The considerations of haplotype frequencies and the way they are reported necessitate the “counting” approach (see Chapters 6 and 7). The strength of the evidence depends on the sizes of the databases. The probability of exclusion, or random man not excluded (RMNE), or the complementary probability of inclusion entails a binary view of alleles, meaning that alleles are only present or absent. Furthermore, if they are present, they are observed. If alleles are found where there is a possibility of stochastic effects, laboratories may omit the inconvenient loci from their calculations (Gill et al., 2006). Such a calculation incorrectly implies that among the “random men” considered for comparison, only the same loci as those considered for the suspect in question would be used for inclusion or exclusion (see Chapter 5 for low level DNA techniques). Two methods of statistical significance were presented in the O.J. Simpson case in California (Weir, 1995). The prosecution wished to use the LR and the defense wanted to use the RMNE. The final result was that the court heard both methods and ruled that the LR method was preferable. Also see Chapter 1 for a discussion of this case. Clayton and Buckleton (2005) summarized the advantages and disadvantages of each approach. Full discussions of the various methods of interpreting evidence can be found in comprehensive texts (Buckleton, 2005; Balding, 2005). According to the DNA Commission of the International Society of Forensic Genetics (Gill et al., 2006), the scientific community has a responsibility to support improvement of standards of scientific reasoning in the courtroom. This implies that concepts such as likelihood ratios, whether difficult to convey or not, are the methods of choice for the statistical evaluation of DNA profiles. Computer software is available to forensic laboratories for calculating statistics such as likelihood ratios. Some laboratories may perform manual calculations to check their results, although the calculations may be very demanding. Each particular laboratory must have validated the population databases and genotype frequencies it uses in forensic calculations.



Chapter four:  Evidential value and statistics



71



4.3.4 Identity and rarity It is important to note that statistical analyses can never lead to absolute conclusions. DNA evidence is essentially probabilistic as shown above and an expert witness should never denote an individual as the donor of a genetic material from which DNA was produced. There is a growing realization that all forensic science evidence is probabilistic and no current forensic technology supports the unique identification of an individual. Other forensic science disciplines follow binary match or ­no-­match systems and this transparency deficit is being addressed (National Research Council, 2009; Fingerprint Inquiry, 2011). Two authors (Saks and Koehler, 2005) described the ­genetics-­based model of DNA profiling as highlighting the deficiencies in other forensic disciplines in which “untested assumptions and s­ emi-­informed guesswork are replaced by a sound scientific framework and justifiable protocols.” The statistics quoted in forensic reports for DNA profiles are often rarer than “1 in 1 trillion,” a number that is greater than the population in the world (currently 6 billion). These statistics appear incredulous to many people and their method of derivation difficult to understand. It is hoped that this text explains that the statistics in most criminal cases are derived according to assumptions made both in the comparison of DNA profiles, and the quality of the profile itself (complete or partial/low level/mixture). It is also the probability of the DNA profile occurring in a particular population, not the probability of the case hypothesis (see Section 4.4 for legal fallacies). An interesting example of how statistics can be readily misinterpreted is the famous (at least in statistical circles) “birthday problem.” This particular problem has been used to illustrate misconceptions in DNA database matches (Weir, 2007; Kaye, 2009). Assume that equal numbers of people are born every day of the year. Then the random match probability for a particular birthday is 1/365. However, there is over a 50% probability that two people in a group of 23 or more share a birthday. How could this be? This is because there are 253 pairs of people in a group of 23 and the particular birthday is not specified. When translated to DNA issues, the birthday problem has to do with multiple occurrences of any profile, not one particular profile (Weir, 2007).



4.4 Legal fallacies Using unfamiliar terminology plus difficulties in statistical interpretation may lead a legal professional to translate results to a wider perspective that may not be valid. Two well-known fallacies are common in the legal community and sometimes even in the news media. The prosecutor’s fallacy is also called the “fallacy of the transposed conditional.” This fallacy



72



Introduction to forensic DNA evidence for criminal justice professionals



translates the chance probability of a crime stain match to the probability of innocence. For example, say there is a 1 in 100,000 chance probability of a match in a city of 1 million people. The prosecution fallacy is to say there is a probability of innocence of 1 in 100,000. The defense fallacy in this particular situation is to say the probability of guilt is 1 in 10. Suppose a crime is committed in London (population about 7 million) and a crime scene profile has a likelihood ratio (LR) of 1 in 1 million. The prosecutor might say that the odds are a million to one in favor of the defendant being guilty. However, based on population size, about seven people in the city are expected to match the profile so it can be argued that the odds are actually 7 to 1 in favor of innocence. The defense fallacy unrealistically assumes that each of the 7 people has equal probability of guilt. An often-quoted case from England (R. v Deen, 1994; P ­ uch-­Solis et al., 2012) illustrates the prosecutor’s fallacy. Deen was an early DNA case in which the random match probability was quoted as 1 in 3 million. Prosecutor: So the likelihood of this being any other man but Andrew Deen is 1 in 3 million? Expert: In 3 million, yes. Prosecutor: You are a scientist … doing this research. At the end of this appeal a jury are going to be asked whether they are sure that it is Andrew Deen who committed this particular rape in relation to Miss W. On the figure which you have established according to your research, the possibility of it being anybody else being 1 in 3 million, what is your conclusion? Expert: My conclusion is that the semen originated from Andrew Deen. Prosecutor: Are you sure of that? Expert: Yes. The basic fallacy is contained in the first question when the attorney asks the probability of the accused being the source of the DNA profile; the attorney should have asked about the probability of the evidence. It is the jury’s responsibility to decide whether factual propositions have been established by the evidence, not the expert. Having been asked the wrong question, the expert in Deen confounded the fallacy, even to the extent of pronouncing himself “sure” that Deen was the source of the stain. In fact, a random match probability of 1 in 3 million implies that about 20 people in the UK would be expected to share the same profile. The prosecution fallacy (transposing the conditional) may be described by two simple statements (Aitken et al., 2010):



1. If I am a monkey, I have two arms and legs. 2. If I have two arms and legs, I am a monkey.



Chapter four:  Evidential value and statistics



73



This logic problem can be avoided by using the LR strictly as quoted in the forensic report. The probability of the evidence based on the hypothesis should not be translated to the probability of the hypothesis itself. It is also helpful to remember that DNA profiling evidence provides only the probability of a match of DNA profiles in the relevant population, not the probability that a particular person committed the crime. As will be repeated throughout this book, DNA is only one piece of evidence in a crime. Limitations of the evidence must be described. The question of how the DNA was transferred is one for the jury to consider. The scientist’s main role is to outline the various modes of transfer that exist and advise on the relative risks associated with the modes (Gill and Buckleton, 2010). The uncertainties about the mode of transfer increase with touch DNA evidence—evidence that cannot be associated with a particular body fluid (Buckleton, 2009).



4.5 Understanding reports: Common phrases and their meanings Identifying the strengths and limitations of facts and opinions is a cornerstone of forensic science. Any forensic report or testimony should convey the limitations of all tests and all the evidence. All conclusions, assumptions made, and inferences should be enunciated and clearly explained. Differences or similarities between evidence and reference samples should be explained as actual differences or similarities inherent in the evidence or as consequences caused by imprecision of the test system— limitations. All alternative explanations (such as different hypotheses proposed) should also be conveyed in the report or testimony.



4.5.1 Inclusion and exclusion Scientific statements should clearly support or refute a finding or state that the result is not possible due to the limitations of the hypotheses proposed. Case 2 from Western Australia (Merritt, 2010) shows how misconceptions may arise from the wording of forensic statements. Case 2 ­ ixteen-­year-old Patrick Waring was accused of rape, spent a S year in detention, and was exonerated in 2007. The forensic report stated that the accused “could not be excluded” from the DNA profile taken from the victim’s underwear.



74



Introduction to forensic DNA evidence for criminal justice professionals Not only is this poor English (double negative) but the scientific use of “excluded” in a forensic laboratory context is unclear to a lay person. A defense expert ­re-­examined the mixed DNA profile and concluded that no DNA evidence linked the accused with the victim.



It can be confusing to the legal practitioner to delineate the excluded, inconclusive, and not excluded terms when the rationale for assigning a specific term to a specific DNA profile is not explained in the report. The terminology is especially problematic when the possible contributing DNA cannot be accorded a statistical weighting, either when it is denoted as inconclusive or not excluded. This implies the evidential value is similar whether a result is inconclusive or not excluded. One strength of DNA evidence is its high discrimination power and thus power to exclude; another is the ability to provide statistics for the probability of inclusion. When one or both of these strengths are absent, the DNA evidence becomes commensurately limited. An illuminating study on an adjudicated criminal case (Dror and Hampikian, 2011; Geddes, 2010) involved a DNA mixture from a gang rape. Two analysts from the original laboratory stated that Suspect 3’s DNA profile “could not be excluded” from the mixture profile. Both profiles were presented to 17 DNA analysts from the same accredited government laboratory without contextual information. The results of the analysts were not consistent. Only one analyst stated that Suspect 3 “cannot be excluded.” Of the remaining 16, 4 stated “inconclusive” and 12 stated “excluded.” The authors of the study suggested subjectivity was present in mixture interpretation. They also note that bias may have originally occurred since the results were not consistent with the original result obtained by analysts who had contextual information. See Chapter 8 for more discussion on this case. If any possible individual contributing DNA to a mixed sample is deemed an inclusion, an associated statistical analysis must support that inclusion. According to the Scientific Working Group on DNA Analysis Methods (SWGDAM) Interpretation Guideline 4.1, a laboratory must perform statistical analysis in support of any inclusion in the context of a case, irrespective of the number of alleles detected and the quantitative value of the statistical analysis (Buckleton et al., 2007; SWGDAM, 2010). Inclusion, included, and cannot be excluded all convey the same meaning and thus statistical frequencies must be reported with any statement that includes them. According to Section 3.6 of the SWGDAM interpretation guidelines, if the known individual “cannot be excluded” from the profile, he or she must be “included” (SWGDAM, 2010). Furthermore, Section 4.3



Chapter four:  Evidential value and statistics



75



states that “the laboratory must not use inconclusive/­unreportable data (e.g., at individual loci or an entire ­multi-­locus profile) in statistical analysis.” Guideline 4.1 states specifically that “the laboratory must perform statistical analysis in support of any inclusion that is determined to be relevant in the context of the case, irrespective of the number of alleles detected and the quantitative value of the statistical analysis.”



4.5.2 Declared contributor If an individual is accepted by all parties as a contributor to the DNA detected, he or she is a declared contributor. Often in sexual offense cases in which an intimate medical swab from a complainant reveals female and male DNA, the complainant is considered a declared contributor to the mixed DNA present. It is important in a criminal case to obtain as many reference DNA samples as required for elimination purposes and for determination of declared contributors. The principles of mixture analysis should be borne in mind for semen stains from clothing and intimate medical samples from complainants in rape cases as shown in Case 3 (Thompson et al., 2003). Case 3 An 11-year-old girl was raped by the pool of her home in Oklahoma in 1991. Timothy Durham was a local resident with a record of criminal violations and the police focused on him. The victim identification and hair comparison evidence was inconclusive but a semen stain on the victim’s swimsuit allegedly matched the DNA (DQ-­alpha) of Durham. Despite 11 alibi witnesses who said he was in a different state at the time of the crime, Durham was convicted in 1993 and sentenced to over 3,000 years in jail. In 1996, he contacted the Innocence Project and asked for further DNA testing of the semen stain. The new tests showed that Durham did not share the ­DQ-­alpha type present in the semen stain and he was also excluded at several other genetic loci. The initial ­DQ-­alpha test was shown to be a false positive because the laboratory failed to separate completely the male and female donor samples during the differential extraction of the semen stain (Section 3.2.3 in Chapter 3 covers differential extraction). The victim’s alleles when combined with those of the true rapist matched the alleles of Durham. The laboratory mistook this mixture for a single source. Durham was released in 1997.



76



Introduction to forensic DNA evidence for criminal justice professionals



When calculating LRs for a ­two-­person mixture, the declared contributor is included in both the defense and prosecution propositions. If a declared contributor can be assumed to be part of a mixture, the issue can be resolved into a single additional contributor.



4.5.3 Verbal descriptors The verbal descriptor scale was devised by Ian Evett and Bruce Weir in England in the late 1990s and later expanded (Evett and Weir, 1998; Buckleton et al., 2005). The scale is commonly used in England and Australia. It is designed to provide an indication of the value of a particular LR, using terminology intended to be consistent between scientists. The scale ranges from extremely strong for the prosecution proposition (LR greater than 1 million) to inconclusive (LR of 1) to extremely strong support for the defense proposition if the LR is less than .000001. Two advantages of DNA profiling (not found with other forensic disciplines) are its high discrimination power and the generation of statistics. This author believes that after a verbal descriptor is applied, the descriptor has a chance of being preferred as a simpler option by a n ­ onscientific reader. This may then reduce what is a comparative analysis of two propositions—inherent to the LR and thus evidential value—to an unwarranted conclusion. The situation is particularly confusing when mixture DNA profiles must be considered.



4.6 Sampling correction and uncertainty A sampling (or size bias) correction is used to allow for the limited database sizes used in frequency calculations. Various methods are used in different jurisdictions. A common method uses the Balding and Nichols equation that adds the case profiles as additional empirical data (Balding and Nichols, 1994). When some laboratories calculate LRs, a sampling uncertainty is also factored into the result. The LR estimates are calculated from a sample of the population. Data from another sample of the same population may yield a different estimate. The method allows determination of the best estimate and a lower value of LR that has a 99% probability of being lower than the true value. This means that there is a 1% probability that the true LR is greater than the value quoted. The true value of the LR can be determined only if sampling uncertainty can be eliminated. Finding a true value would require DNA typing of an entire population which is currently not possible.



Chapter four:  Evidential value and statistics



77



4.7 Relevant population and impact on statistical value When determining LRs, two alternative propositions must be considered: (1) the defense proposition is that the person of interest is not the donor and (2) someone else contributed the DNA. In many cases the “someone else” is of an unknown ethnic or racial origin. However, the defense may wish to state (or the parties may agree to) designation of a particular ethnic source of the suspect, e.g., Victorian Caucasian. The most populous databases in Western countries such as the United Kingdom, the United States, and Australia are Caucasian. Calculations for other ethnic groups can be performed if the defense or other party desires. Some ethnic groups may still be isolated genetically and culturally or essentially incorporated into the general population due to current society norms. Consequently it is important to ensure that the statistical calculations are performed on the relevant population. An example of new ideas surrounding population databases is the segregation of the Australian aboriginal population data along contemporary state and territory lines. This appears to mask the diversity that exists within this subpopulation. Datasets collected among more traditional lines may be more appropriate, particularly to distinguish the most genetically differentiated populations residing in the north of the continent (Goetz et al., 2008; Walsh et al., 2007). The ­co-­ancestry coefficient addresses the fact that two people within an ethnic group are more likely to have similar genotypes than two people from different ethnic groups. This is represented as Fst or theta. The use of general population frequencies and the product rule will disadvantage a suspect because the general population may not necessarily display the same frequencies as the subpopulation (Buckleton, 2005).



4.8 Relatives Further analytical work must be performed if the suggestion is made that a blood relative (e.g., a brother) of the accused is the true perpetrator. This is the ­so-­called brother defense. If it is possible to obtain a reference sample from the relative and full DNA profiles have been developed, they can be compared and thus reduce the need for further statistical work. If a reference sample is not obtainable, calculations based on the relatedness are required. Mendel’s theory of heredity (see Chapter 1) proposes that a parent is equally likely to pass on either of their two alleles to offspring.



78



Introduction to forensic DNA evidence for criminal justice professionals



The inbreeding coefficient Fst is 0.25 for siblings and 0.0625 for cousins, and the recognized reference (Balding and Nichols, 1994) has the values for the most common relationships. Many forensic laboratories have packages that calculate the statistics for relatives but they are applicable only to single source profiles. SWGDAM Guideline 5.2.3 (2010) covers relatives.



References Aitken, C., Roberts, P. and Jackson, G. 2010. Practitioner Guide 1: Fundamentals of Probability and Statistical Evidence in Criminal Proceedings. London: Royal Statistical Society (available online). Ayturgrul v. The Queen. High Court of Australia, HCA 15, April 18, 2012. Balding, D.J. 2005. Weight of Evidence for Forensic DNA Profiles. Chichester: John Wiley & Sons. Balding, D.J. and Nichols, R.A. 1994. DNA profile match probability calculation: How to allow for population stratification, relatedness, database selection, and single bands. Forensic Science International, 64, 125, 1994. Buckleton, J. 2005. A framework for interpreting evidence. In Buckleton, J. et al., Eds., Forensic DNA Evidence Interpretation. Boca Raton, FL: CRC Press, 27–63. Buckleton, J. 2009. Validation issues around typing of low level DNA. Forensic Science International: Genetics, 3, 225–260. Buckleton, J., Curran, J., and Gill, P. 2007. Toward understanding the effects of uncertainty in the number of contributors to DNA stains. Forensic Science International: Genetics, 1, 20–28. Buckleton, J., Triggs, C.M., and Walsh, C.J., Eds. 2005. Forensic DNA Evidence Interpretation. Boca Raton, FL: CRC Press. Chakraborty, R. 1992. Sample size requirements for addressing the population genetic issues of forensic use of DNA typing. Human Biology, 64, 141–159. Clayton, T. and Buckleton, J. 2005. Mixtures. In Buckleton, J. et al., Eds., Forensic DNA Evidence Interpretation. Boca Raton, FL: CRC Press. Curran, J. and Buckleton, J. 2011. An investigation into the performance of methods for adjusting for sampling uncertainty in DNA likelihood ratio calculations. Forensic Science International: Genetics, 5, 512. Dror, I.E. and Hampikian, G. 2011. Subjectivity and bias in forensic DNA mixture interpretation. Science and Justice, 51, 204–208. Evett, I.W. and Weir, B.S. 1998. Interpreting DNA Evidence, Statistical Genetics for Forensic Scientists. Sunderland, MA: Sinauer. Fingerprint Inquiry. 2011. Inquiry Report. Edinburgh. http://www. thefingerprintinquiryscotland.org.uk/inquiry/3041.html Geddes, L. 2010. Fallible DNA evidence can mean prison or freedom. New Scientist, August. http://www.newscientist.com/article/mg20727733.500 Gill, P. 2001. Low copy number DNA profiling. Croatian Medical Journal, 42, 229–232. Gill, P. 2002. Role of short tandem repeat DNA in forensic casework in the UK: Past, present and future perspectives. BioTechniques, 32, 366–385. Gill, P. and Buckleton, J. 2010. A universal strategy to interpret DNA profiles that does not require a definition of low copy number. Forensic Science International: Genetics., 4, 221–227.



Chapter four:  Evidential value and statistics



79



Gill, P., Brenner, C.H., Buckleton, J.S. et al. 2006. DNA Commission of the International Society of Forensic Genetics: Recommendations on the interpretation of mixtures. Forensic Science International, 160, 90–101. Gill, P., Whitaker, J., Flaxman, C. et al. 2000. An investigation of the rigor of interpretation rules for STRs derived from less than 100 pg of DNA. Forensic Science International, 112, 17–40. Goetz, R., West, J., Walsh, S.J. et al. 2008. Population data from the NSW Aboriginal Australian ­sub-­population for the Profiler Plus autosomal short tandem repeat (STR) loci. Forensic Science International, 175, 235–237. Kaye, D. 2009. Trawling DNA databases for partial matches: What is the FBI afraid of? Cornell Journal of Law and Public Policy, 19, 145–171. Merritt, C. 2010. Film shines a light on the dangers of DNA dependence. The Australian, May 21. National Research Council. 2006. The Evaluation of Forensic DNA Evidence. Washington: National Academies Press, p. 130. National Research Council. 2009. Strengthening Forensic Science in the United States: A Path Forward. Washington: National Academies Press. ­Puch-­Solis, R., Roberts, P., Pope, S. et al. 2012. Practitioner Guide 2: Assessing the Probative Value of DNA Evidence. London: Royal Statistical Society (available online). R. v Deen, Court of Appeals, U.K. The Times, January 10, 1994. Saks, M.J. and Koehler, J.J. 2005. The coming paradigm shift in forensic identification science. Science, 309, 892. SWGDAM (Scientific Working Group on DNA Analysis Methods). 2010. Interpretation Guidelines for Autosomal STR Typing by Forensic DNA Testing Laboratories. FBI website: http://www.fbi.gov/about-­us/lab/codcdis/swgdam.pdf Thompson, W.C. 2003. Review of DNA evidence in State of Texas v. Josiah Sutton. February 6. http://www.scientific.org/archive/Thompson%20Report.pdf U.S. Federal Judicial Center. 2011. Reference Manual on Scientific Evidence, 3rd ed. Washington: National Academies Press. Walsh S.J., Mitchell, R.J., Curran, J.M. et al. 2006. The extent of substructure in the indigenous Australian aboriginal population and its impact on DNA evidence interpretation. International Congress Series, 1288, 382–384 Weir, B.S. 1995. DNA statistics in the Simpson matter. Nature Genetics, 11, 365–368. Weir, B. 2007. The rarity of DNA profiles. Annals Applied Statistics, 1(2), 358-370.



chapter five



Partial profiles, low levels, and mixtures DNA profiling is often considered by the lay public to generate complete profiles, but crime scene profiles are not produced by pristine reference samples. The result may be partial and degraded, at low level, or consist of mixtures from multiple contributors. Special considerations exist when a DNA profile is not complete or comes from more than one source. The legal practitioner should be alert to the issues explained below. A lower-than-expected statistical value that implicates a client (for example, less than a default value typically in the billions) should always be investigated further. Report notations indicating a partial profile, mixture, low level, or trace amount should also be explored further.



5.1 Partial profiles A full DNA profile is always the aim of profiling analysis. However, an incomplete or partial DNA profile may be obtained due to degradation of the DNA in the sample, insufficient quantity or quality of the sample, or a combination of these factors. A statistical evaluation involving a partial DNA profile will generally have less statistical value than a complete profile. The Crown Prosecution Service in England reported that half the DNA profiles yielded from samples recovered from crime scenes are partial profiles (2010). A partial DNA profile may indicate low levels of DNA. When a sample contains small amounts of DNA, the larger fragments of DNA may fail to amplify; only the smaller fragments amplify and the results at the higher end of the electropherogram are missing. A partial DNA profile may also indicate degradation or inhibition of the DNA. Degradation of a DNA molecule occurs over time, particularly when subjected to heat, sunlight, water, and/­or bacteria (see Chapter 3). Thus it is not uncommon to see partial DNA profiles in cold cases or from outdoor crime scenes. Degradation is often signaled by the “ski slope” effect whereby the heights of the alleles decline toward the right of the graph in the 81



82



Introduction to forensic DNA evidence for criminal justice professionals



electropherogram, like going from the top to the bottom on a mountain or ski slope. Sometimes the alleles toward the right side of the graph disappear. This can also be the situ­ation with inhibited DNA (Chapter 3). The alleles toward the left side of the graph (smaller DNA fragments) may still generate good height without being considered low level. Thus it may be prudent to have a DNA analyst explain a particular profile in detail. Figure  3.5 (Chapter  3) and Figure  5.1 are both partial DNA profiles exhibiting degradation or inhibition of DNA molecules.



5.2 Low level and suboptimal profiles There has been increasing debate in the forensic science and legal communities regarding the analysis, interpretation, and meaning in cases of low amounts of DNA. Forensic DNA is a complex field of science and the debates have not been resolved. Every report and laboratory should acknowledge the complex meanings of trace or low level DNA, define the parameters in which the results may be interpreted with confidence, and communicate these limitations to the criminal judicial system.



5.2.1 Definitions There is a proposal that there should not be a definitive line between what is considered low level DNA and conventional autosomal STR typing (Gill and Buckleton, 2010). The literature cites several definitions: • Amount of DNA tested in the PCR reaction (for example, less than 200 picograms) based on assay quantification • Increasing the number of PCR cycles beyond 28 • DNA profile appearance that exhibits stochastic effects (see below) Low level DNA is now generally accepted as applying to situations in which the amount of DNA available is limited or interpreting the resulting profile may require more considerations than profiles generated using higher amounts of DNA. Low level DNA may also be defined as any sample that falls below recommended thresholds at any stage of the analysis, from sample detection through to profile interpretation, and cannot be defined by a precise picogram (extremely small weight, abbreviated as pg) level (van Oorshot et al., 2010). These profiles are considered suboptimal and should be designated as such in any report or testimony. Data reliability is inferior and thus additional measures must be taken to accurately reflect the sample examined (Butler, 2012). Until the past decade or so, DNA was recovered only from visible stains such as blood splatters and semen stains. Now technology has



100



110



120



130



140



vWA



150



160



170



180



190



200



210



220



D16S539



230



240



250



260



270



280



D2S1338



290



300



310



320



330



340



350



360



120 80 40 0



240



15



100



AM...



110



D8S1179



120



17



130



17



140



150



160



170



180



19 D21S11



190



11 12



OL OL



200



210



D18S51



220



230



240



250



260



270



280



290



300



310



320



330



340



350



360



220



230



240



250



260



270



280



290



300



310



320



330



340



350



360



180 120 60 0



X



Y



11



D19S433



100



110



120



130



140



14 150



28 TH01



160



170



180



190



200



FGA



210



400 300 200



Chapter five:  Partial profiles, low levels, and mixtures



D3S1358



160



100 0



12



14



6



23



Figure 5.1  Low level partial DNA electropherogram showing degradation and/­or inhibition of DNA and considerable ­drop-­out. All peaks except the X peak are below 200 RFU (inferring that they are below the stochastic threshold). OL = o ­ ff-­ladder peak.



83



84



Introduction to forensic DNA evidence for criminal justice professionals



advanced and samples are collected (usually by swabbing) from areas that exhibit no visible staining but might be expected to reveal biological matter deposited through handling, for example, knives and steering wheels. Very small amounts of DNA (fewer than 100 pg or 0.0000000001 g) may produce a DNA profile. But how far do we push these DNA testing techniques so that we can be confident of reliable results? The amplification kits commonly used in forensic laboratories usually recommend a sensitivity threshold of at least 250 pg of template DNA. They are not validated for quantities of DNA below that very small amount. Issues surrounding the interpretation of DNA profiles using low level analytical techniques such as low copy number (LCN) were brought to the attention of the scientific community and the public domain after this type of evidence was questioned by the presiding judge in the Omagh bombing appeal in Belfast 2007 (Case 1). The accused was freed and LCN DNA use was suspended in British courts for a period. Case 1 The Omagh bombing occurred in 1998 and 29 people were killed and 220 injured in a car bomb attack. Sean Hoey was charged in 2005 after it was alleged that his DNA was found on bomb timers collected through crime scene examination. However, the LCN technique did not exist in 1998 and crimescene examiners did not necessarily follow the stringent ­anti-­contamination requirements for such examinations. An independent report (Caddy et al., 2008) found that the laboratory methods were robust and validated but confusion remained in the interpretation of such profiles. The report recommended that a DNA profile using low template DNA techniques should be presented with clear caveats to juries in criminal trials. What is categorized as low level DNA may also be called low copy number or low template DNA. What is important is that the laboratory or the report defines the term used to describe low amounts of DNA and explains the type of enhancement technique (such as increased amplification) used, if any. The definitions should be stated in the body of the report and/­or any appendices. Gill et al. (2000) suggest insertion of a clause in expert statements cautioning the court on the lack of interpretative information such as transfer and persistence studies when determining the value of low level DNA.



Chapter five:  Partial profiles, low levels, and mixtures



85



While many laboratories perform testing with small amounts of DNA, only a few formally conduct low template DNA case work with specific enhanced detection protocols (Butler, 2012). There has recently been a review commissioned by the Home Office in the United Kingdom to address principles involved in the interpretation of DNA profiles, especially those that are complex in some way because the target material is at a low level, or degraded (Gill et al., 2012). The review noted that the strength of evidence (to support a prosecution or defense hypothesis) is likely to be maximized with the full conventional DNA profile and minimized with the poorest interpretable low level DNA profile.



5.2.2 Stochastic effects A forensic laboratory needs to determine at what point a detection technique cannot deliver reliable results. Stochastic effects are random sampling effects that may occur when a limited number of DNA target molecules exist in a sample. Stochastic is derived from the Greek language and describes systems whose behaviors are intrinsically ­nondeterministic or random. What happens with low amounts of DNA is that the PCR primers used to amplify a specific region may not consistently find and hybridize to the entire set of DNA molecules present in the amplification reaction. With a heterozygous locus in which two alleles are present, unequal sampling of the alleles can result in failure to detect one or both alleles. Loss of a single allele is called “­drop-­out” while loss of both alleles is termed “locus ­drop-­out.” Other effects are unbalanced peak heights of paired alleles and masking from a known or unknown contributor. ­Drop-­out arises when the allele carried by an individual contributing to a sample is not reported within the DNA profile obtained from the sample. ­Drop-­in occurs when trace amounts of DNA, for example, from the crime scene environment or laboratory plasticware, generate one or more spurious alleles in a profile. It is rare for ­drop-­out and d ­ rop-­in to occur with good quality samples not subjected to degradation or inhibition, but d ­ rop-­in and d ­ rop-­out become more likely as DNA amounts decrease or environmental exposure increases. Figure  5.1 shows both allele and whole locus ­drop-­outs. Complete locus d ­ rop-­out appears at D2S and at D18. This electropherogram may be described as a partial DNA profile in a forensic report. All peak heights appear to be below 200 RFU except for the X sex marker. In the author’s opinion this is a low level, suboptimal DNA profile that should not be



86



Introduction to forensic DNA evidence for criminal justice professionals



used for comparison purposes unless further analyses are performed to obtain a better profile. Stochastic variation is a fundamental physical law of the PCR amplification process when low amounts of DNA are examined. It manifests as a fluctuation of results between replicate analyses. Thus it is possible that amplifying the same extract twice can result in detection of different alleles at a locus (Butler, 2012). Figure 5.2 shows two electropherograms of the same sample run twice. Note the different peaks at the D2 locus. Since stochastic effects cannot be avoided when testing small amounts of DNA, two approaches have been proposed: (1) stop testing or interpreting data before the stochastic realm is reached, or (2) try to limit the impact by performing additional testing and following careful interpretation guidelines based on validation studies. The second, or “enhanced interrogation,” approach typically involves replicate testing and the development of consensus profiles (Butler and Hill, 2010; Butler, 2012). Those who advocate the second approach usually enhance method sensitivity in order to get as much yield from a sample as possible. Careful validation studies and appropriate interpretation guidelines are essential. Recently, an approach has been advocated to completely eliminate template thresholds from the definition because they represent an artificial ­cut-­off for a continuous phenomenon. Instead, a risk assessment based on peak heights of the DNA profile can be used to determine whether an appropriate amount of DNA is present and stochastic factors are impacting the typing (Gill and Buckleton, 2010). This paper describes a statistical package that accounts for d ­ rop-­out and d ­ rop-­in in probability terms. This paper also critically examines the causes of the underlying confusion about low template and low level DNA profile interpretations. The biological model with replicate and consensus profiles was designed to prevent misstatements about the strength of the evidence and present suitable warnings about the limitations. The authors state that candor in reporting should be applied. They believe that their statistical model is the way forward, but in the absence of validated software implemented in the specific laboratory issuing the report, the biological model can be used with limitations expressed. This approach has also been recommended by the Home Office (Gill et al., 2012). One of their principles is that DNA interpretation methodology should incorporate a probabilistic consideration of drop-out and additional alleles, such as drop-in, stutters, gross contamination and additional contributors. Fraser review—In 2010, in Victoria, Australia, a group of international experts recommended changes in the state laboratory’s DNA profile interpretation methodologies. The chief commissioner of police suspended the presentation of DNA evidence in court for six weeks; low level DNA profiles were of particular concern (Fraser et al., 2010).



vWA 150



D16S539 200



250



D2S1338 300



350



63 42 21 0



OL



17



16



10



OL



24 26



11



D3S1358 100



D16S539



vWA 150



200



250



D2S1338 300



350



150 100



Chapter five:  Partial profiles, low levels, and mixtures



D3S1358 100



50 0



OL



17



16



10 11



24



87



Figure 5.2  Example of stochastic effects. R ­ e-­analysis of the same sample revealed dfferent results at locus D2.



23



88



Introduction to forensic DNA evidence for criminal justice professionals



This action was taken due to the growing unease of the police and the criminal justice community about the laboratory’s methodologies. Three experts from the United Kingdom and New Zealand were invited to undertake a review of the laboratory’s DNA interpretation practices. The team was headed by Professor Jim Fraser and included John Buckleton and Peter Gill, authors of numerous papers described in this book. The review considered that the main events precipitating the concern appeared to be (1) variations in statistics upon the implementation of a new method designed to deal with peak ­drop-­out due to low level or degraded DNA, (2) dealing with artifacts in the profiles, and (3) inconsistencies in interpretations of profiles by forensic scientists in the laboratory. Very different statistical outcomes resulted from different methodologies in different situations, and the differing outcomes raised concerns by the police and prosecutors and generally resulted in a loss of confidence. The Fraser team considered that removing professional judgment from case managers was misguided because DNA profiling cannot be deskilled to such an extent. They considered that the concern and loss of confidence also arose from the organizational environment in the laboratory. The physical conditions were in need of significant improvement and the cramped conditions and proximity of samples presented unnecessary risks of contamination. The experts recommended the development of a better appreciation of DNA interpretation practices throughout Australia and internationally, broader engagement of staff, and the implementation of a new DNA interpretation method. The findings of the Fraser review can be applied to the consideration of DNA evidence by criminal justice professionals in any jurisdiction. Problems in interpretation of low level DNA profiles are ­drop-­in and ­drop-­out, stutter, and unbalanced peak heights, in addition to masking from a known contributor. There is a developed framework described above for assessing such evidence based on likelihood ratios (LRs) that involve ­drop-­in and d ­ rop-­out probabilities (Gill and Buckleton, 2010). Ignoring a discrepant locus or using the “random man not excluded” approach can be systematically unfair to defendants. The LR depends strongly on the assumed probability for ­drop-­out, and ignoring the possibility of ­drop-­in is unfair to defendants.



5.2.3 An interesting experiment A criminology professor in the United States performed an experiment involving inadvertent participants (DNA analysts) attending meetings and conferences over a number of years (Thompson, 2009). They were shown part of the electropherogram (Figure 5.3) of an evidentiary sample (presumed to be saliva) swabbed from the skin of a sexual assault victim.



Chapter five:  Partial profiles, low levels, and mixtures D3S1358



vWA



89



FGA 100 50



12



17



15



40



94



56



OL Allele ? 84 17 90



Defendant D3S1358 Tom 17,17 Dick 12,17 Harry 14,17 Sally 12,17



vWA 15,17 15,17 15,17 15,15



25



OL Allele ? 90



143



FGA 25,25 20,25 20,25 20,22



Figure 5.3  Electropherogram of saliva sample and four suspect profiles. (Source: Thompson, W.C. 2009. Law, Probability, and Risk, 8, 257–276. With permission.)



The table at the bottom of the figure shows the alleles of four possible defendants. Thompson stated that the choice of defendants who should have been included or excluded as possible contributors to the evidentiary sample was unclear. At locus D3S1358 (D3) it must be determined whether the peak labeled 12 represents a true allele and, if so, whether it is associated with (from the same contributor as) allele 17. At locus FGA, the determination must be made whether the peak labeled “OL allele?” is a true allele or an artifact. Another consideration is whether the electropherogram represents a single source sample or a mixture. At the first meeting, Thompson noted some uncertainty about the inclusion of Defendant Tom who did not have the 12 peak at locus D3 and asked how the participants could be sure that the true contributor did not have genotype 12,17 at locus D3. Several analysts argued that the 12 peak at D3 and the OL artifact should have been ignored because of the height disparity between the 12 and 17 alleles and the poor morphology of the 12 allele. Of course, Defendant Tom was included. At the next meeting, Thompson presented the evidentiary profile and the defendant was presented as Dick rather than Tom. Thompson suggested that the inclusion of Dick was problematic because of the uncertainty whether the 12 peak at D3 was a true allele and because no 20 peak had been detected at locus FGA. One analyst said the peak height disparity was not an issue because these discrepancies are expected due to



90



Introduction to forensic DNA evidence for criminal justice professionals



stochastic effects. The OL allele was indeed an artifact that could have masked an underlying 20 allele. At a subsequent meeting, Thompson presented the defendant as Harry. The analysts found no problem with this inclusion as the failure to detect the defendant’s allele 14 at locus D3 could easily be due to allelic drop out and a 20 peak at locus FGA may have been masked by an artifact. The peak labeled 12 at locus D3 was an obvious artifact. At that point, Thompson wondered how much he needed to change the defendant profile to get the analysts to agree that the defendant should have been excluded. A colleague of Thompson’s, Dan Krane, also presented the case but this time using a defendant profile labeled Sally. The analysts still insisted that the defendant could not be excluded and invoked a mixture of two contributors, one of whom had the 15 allele at locus vWA and the other which had the 17 allele (Thompson, 2009). The problems with this electropherogram are (1) it is a low level DNA profile with all peaks below any laboratory stochastic threshold, and (2) it presents the possibilities of d ­ rop-­out, ­drop-­in, and artifacts. In the opinion of this author, this type of profile is suboptimal and not suitable for comparison with reference samples. It is far better to ­re-­analyze the DNA extract to obtain a better quality profile. If this is not possible, the electropherogram should be marked as “not interpretable due to quality of profile.”



5.2.4 Enhancement techniques Strategies exist for improving sensitivity in DNA analyses. An increased number of PCR amplification cycles were first described in the late 1990s, but they presented the possibilities of allele ­drop-­out and ­drop-­in and increased risks of collection- and ­laboratory-­based contamination. Later methods increased the sensitivity of the injected product, improved ­post-­PCR purification, and reduced PCR volume (Butler, 2012).



5.2.5 Improving reliability of results There are two main areas where scientists aim to improve the reliability of results obtained when potentially working with low amounts of DNA (such as touch DNA from handled objects). First, they try to improve the amount of DNA recovered at the collection and extraction stages. Second, they limit the possibility of obtaining an incorrect answer by accounting for stochastic effects in the analysis with a biological or statistical model.



Chapter five:  Partial profiles, low levels, and mixtures



91



5.2.5.1 Biological (consensus) model The moment of acceptance of stochastic effects will lead to difficulty in inferring the contributing profiles of individuals from original or replicate electropherograms. The biological model determines the profile via a consensus strategy. The model was originally developed to facilitate the reporting of DNA profiles that were subject to the phenomena of ­drop-­out and ­drop-­in. It was later improved in order to facilitate reporting of low template DNA profiles; the development of software solutions came later. The statistical model discussed in the next section was concurrently made available to check calculations provided by the biological model. The biological model was devised to prevent misstatements of the strength of the evidence and provide suitable caveats and warnings about the limitations to be applied (Gill and Buckleton, 2010). The replicate amplification strategy became the core feature of low level DNA profiling (Butler, 2012). This concept was first published in 1996 (Taberlet et al., 1996). An allele was recorded only if it was observed at least twice. The standard practice to date is to amplify two or three aliquots of a DNA extract (Gill et al., 2000; Caragine et al., 2009). Alleles that occur more than once in the obtained profiles are designated. However, amplification results from a single test may be unreliable due to the stochastic effects (Butler, 2012), as noted above. The Italian appellate court in the case of Amanda Knox and Raffaele Sollecito (Case 7, Chapter 2) held that the failure to perform two amplifications from the blade of the alleged weapon, despite the low quantity of DNA, may be acceptable for initial investigative purposes “… but cannot be accepted when the genetic tests form the basis for evidence of guilt beyond any reasonable doubt” (Hellmann, 2011). The review commissioned by the Home Office (Gill et al., 2012) stated that replication of a test for a compromised sample, although recommended, was not compulsory—provided that the interpretation can be supported by a suitable statistical analysis. Splitting the sample into two parts may be detrimental to interpretation and maximizing the sample size for amplification will reduce ambiguity inherent in the DNA profile of a compromised sample. A single analysis may provide the difference between a test result that can be reported and one that cannot.



5.2.5.2 Statistical (probabilistic) model A strategy for interpreting low level DNA profiles and accounting for stochastic variability was first introduced as a “statistical model” (Gill et al., 2000). The statistical model can be used in two different ways: (1) to develop a likelihood ratio per se or (2) to determine whether the consensus



92



Introduction to forensic DNA evidence for criminal justice professionals



approach is safe under the circumstances described. It attempts to assess the probability of the replicates from all possible genotypes (Gill and Buckleton, 2010). The test of the strength of the evidence is assessed on a continuous basis to formulate the LR. If alleles do not appear or are visualized just a few times in multiple replicates, the LR is low. The emerging literature describes the need for a move by the forensic community toward more formal probabilistic models. The International Society of Forensic Genetics (ISFG) commission supports the adoption of these statistical type models (Gill et al., 2012). However, the slow adoption of these models has been caused by complex concepts that are not routinely encountered in forensic science casework—and are also difficult to describe in a courtroom. The essential feature of the classical or biological model is the prosecution hypothesis of the binary determination of a match versus a ­nonmatch that results in a probability of 1 or 0, respectively. However, with the probabilistic model, the likelihood of the numerator can take any value between 0 and 1 (Gill et al., 2012). Therefore the probability can be described as a continuum, and this constitutes the fundamental difference in the two approaches. If the DNA profiles only partially match between the crime and the reference samples, uncertainty about the validity of the match is present and the numerator cannot be described as 1. Some s­ hort-­cut calculations are used on occasion. The most common is the “2p rule” (Buckleton and Triggs, 2006). This rule is a shortcut method to interpret partial profiles where drop-out has been invoked. For example, only one allele peak may be seen at a locus in the crime profile but the prosecution may invoke drop-out of a partner allele to accord with their hypothesis. The weakness of this rule is that it does not take into account the uncertainty of the match in the numerator of the prosecution hypothesis: is there really a probability of 1 for a match with the suspect profile? It has been shown to be not conservative especially in the presence of “masking” (Balding and Buckleton, 2009). The statistical or probabilistic approach still requires a proper assessment of the overall quality of the DNA profile in question and its suitability for further analysis.



5.2.6 Contamination Contamination is a major issue when considering low level DNA. It is possible to amplify the contaminant through enhancement techniques,



Chapter five:  Partial profiles, low levels, and mixtures



93



so it is imperative that a laboratory employs strict contamination mitigation measures. Crime scene collection procedures should also be subject to scrutiny. Contamination was a major factor in the Omagh bombing appeal and police collection procedures were criticized. Similarly, the collection practices utilized to solve the murder of Meredith Kercher were problematic (Case 7, Chapter 2). The bra clasp from the deceased was collected over a month and a half after the crime and after the house had been subjected to several searches by ­nonscientific examiners in the belief that, by then, all items of scientific relevance had already been found. A photograph of the gloves of the scientific police operatives taken when the clasp was collected showed traces of dirt on the fingertips holding the clasp that could have been interpreted as signs of previous soiling (Hellmann, 2011). The appeal court accepted the theory of probable contamination of the bra clasp. Every forensic laboratory today faces the question of low template or low level DNA and must determine their validated guidelines. An abandonment of DNA testing at low levels is not generally considered practical by the scientific community but a warning of the dangers of not understanding the potential for honest error and margins of error is vital. All forensic scientists should communicate the limitations of their methods to the criminal justice system. The limitations are not confined to low template DNA profiles. A false sense of security is a likely consequence when techniques are represented by artificial divisions, that is, the results obtained from conventional profiles are interpreted using methods that do not follow the same cautions applicable to low template DNA profiles (Gill and Buckleton, 2010). For example, the “Phantom of Heilbronn” case in Germany (Himmelreich, 2009) involved widespread contamination that occurred in relation to conventional DNA profiling (see Chapter 8).



5.3 DNA mixtures from two or more people It is not unusual to observe DNA profiles containing contributions from two or more people. These results may be expected, for example, from vaginal swabs from rape victims that carry mixtures composed of semen from the rapist and vaginal cells from the victim. These types of samples can be readily separated using different extraction techniques for sperm and for skin cells. The problem in interpretation occurs most often with a touch DNA mixture (because skin cells cannot be separated) or when the body origins of the cells cannot be determined. Case 2 from the author’s files is an interesting example.



94



Introduction to forensic DNA evidence for criminal justice professionals



Case 2 A rape case was set for trial. The complainant in the matter alleged she was asleep in the communal lounge room of a boarding house when she awoke to find that the accused (who also lived in the house) had raped her and she saw a used condom on the floor. She said her tracksuit pants and underpants had been removed while she slept. The prosecution relied on the alleged finding of DNA from the accused in a sample from the inner leg of the complainant’s tracksuit pants and maintained that the sample was indicative of semen. No semen was detected on any medical swabs from the complainant. A review of the case notes showed that samples from the left leg near the hem and from the inner knee area of the tracksuit pants reacted positively to a screening test for semen. However, semen could not be confirmed from either sample and in fact the screening test showed slow reaction times for both samples, indicating the possibility of false positives. A major DNA profile was obtained from a ­nonsperm fraction from the hem area that matched the DNA of the accused; no DNA was obtained from the sperm fraction. A mixture DNA profile of at least three people was obtained from the sperm fraction of the inner knee area including a partial DNA profile that matched the corresponding components in the profile of the accused with a statistic of 340,000 to 1.4 million. A method of separating sperm cells from ­nonsperm cells such as cellular material (for example the differential extraction method used in this case) does not confirm the presence of sperm. The hearing transcript also showed that the forensic scientist in the case stated that a sperm fraction in a cellular separation did not confirm the presence of semen. No contribution of the complainant was found on the clothing, and in fact a DNA profile matched half of that of the complainant (possibly from a parent). Further testing still could not confirm semen and the prosecution decided not to lead any DNA evidence in the trial. The differential extraction process for semen and sperm fractions is explained in Chapter 3. Case 2 emphasizes the importance of determining whether DNA can be related to a specific body fluid, also described in Chapter 3.



Chapter five:  Partial profiles, low levels, and mixtures



95



The larger the number of contributors, the more complex the DNA profile. If a sample contains material from four or more contributors, a large portion of the population generally would be included in the profile. Most laboratories do not attempt to perform interpretations of four or more contributors to a mixed DNA profile unless a major contributor can be determined. As described previously, the peaks in DNA profiles represent alleles that are genetically inherited from each parent. When the alleles from each parent are of different sizes, the alleles will have different values. Sometimes alleles of both parents are of the same size and the resulting peak has the same values at each site (effectively doubling the height). When mixtures contain DNA from two or more people, the assignment of each allele to a particular individual becomes more difficult. Computer programs are required to analyze a mixture of two or more contributors without a clear major component. When considering mixture DNA profiles, it is important to consider the profile as a whole to determine the relative quality, quantity, and number of contributors to the DNA. Knowledge of laboratory criteria is required for understanding the assignment of the number of contributors, the designation of a peak in a profile, and the conditions under which it is appropriate to determine a statistic denoting evidential significance. Case 3 illustrates what can happen if defense lawyers accept laboratory reports and expert testimony at face value without examining the underlying scientific data (Thompson, 2003; Bromwich, 2007). Case 3 In 2002, an inquiry by local television reporters into the operation of the Houston (Texas) police crime laboratory led to reviews of several past cases by experts. One of these cases was the conviction of Josiah Sutton for a 1998 abduction and rape committed when he was just 16 years old. A woman was abducted and raped at gunpoint in the backseat of her car and then dumped in a nearby field. She identified Sutton and his friend as her two attackers. The DNA results from the Houston laboratory noted a mixture of DNA from two men, one of whom was Sutton. His friend was excluded. DNA was the primary evidence against Sutton. However, a defense expert found that although the laboratory deemed Sutton’s DNA profile “consistent” with a mixture of alleles found in some samples, the samples contained so many alleles that thousands of people would also be “consistent.” The defense expert excluded Sutton from contributing to the semen



96



Introduction to forensic DNA evidence for criminal justice professionals stain on the backseat. If the unknown man’s DNA on the backseat was in the DNA of the vaginal sample, Sutton could not have been the other man. The jury never heard these possibilities and took only two hours to convict Sutton. What the jury heard was that every human being has a unique DNA pattern and that Sutton’s pattern was found in the sperm fraction of the vaginal swabs, on debris from combing of the complainant’s pubic hair, and on her jeans. The forensic report said that Sutton’s DNA profile would be expected to occur in 1 in 694,000 members of the relevant population. What the jury did not hear was the actual match of the profile in the mixture (approximately 1 in 14). Apparently no mixture statistic for Sutton was given. The inquiry led to DNA retesting of the victim’s vaginal swab by another laboratory. The laboratory found that the semen came from two men and eliminated Sutton. Serious flaws, including incorrect and misleading statistical calculations, were discovered in the original analysis. The DNA section of the Houston crime laboratory was shut down in 2002, reopened in 2006, and again shut down in 2008 and also subjected to audit and review. Sutton was exonerated in 2004 after serving more than 4 years of a 25-year sentence. The real perpetrator was found in 2006 through a DNA CODIS database match. He pleaded guilty and was sentenced to 10 years. The other perpetrator apparently died during imprisonment for other charges.



The interpretation of a mixed DNA profile is relatively simple when the following criteria apply (Word, 2011): • The DNA comes from only two sources. • The two sources are unrelated and have no or few shared alleles. • The ratio of the amount of DNA contributed by each of the two sources is adequate for interpretation of both sources. • The appropriate amount of DNA was amplified and the alleles for both sources exceed the analytical threshold of the laboratory. • No degradation, inhibition, or primer variants are present to affect peak heights and the apparent DNA ratio. • All stutter peaks and other artifacts are below the analytical threshold or clearly distinguishable as artifacts. Significant alteration to any of the above parameters will likely make mixture interpretation more complex; a combination of several alterations generally confounds an interpretation significantly (Word, 2011).



Chapter five:  Partial profiles, low levels, and mixtures



97



According to Guideline 3.6.1 of the Scientific Working Group on DNA Analysis Methods (SWGDAM, 2010; Mixture Interpretation Workshop, 2011), an analyzing laboratory must establish guidelines to ensure that, to the extent possible, DNA typing results from evidentiary samples are interpreted before comparison with any known samples other than those of assumed contributors. Masking of peaks or sharing of alleles is a common result from mixtures because a contributor reference DNA profile from a mixture may have a particular allele or value at a particular locus that is the same as that from another contributor profile. The contributions may be additive if they are similar in amount or they may not be observed if one contributor has donated less DNA than the other(s). It is necessary to make an assessment in relation to the heterozygote balance and mixture proportion (Gill et al., 2006). In some cases DNA from one contributor to a mixed DNA profile may be present in a larger amount than DNA from another contributor. This component is sometimes referred to as originating from the major contributor and may be interpreted as a single source profile. In other cases, none of the contributors to a mixed DNA profile can be inferred (Kelly et. al, 2012). Laboratories have protocols for determining whether major and minor contributions are present in a mixture—often when the minor contribution is less than 30% of the total. The calculations are then relatively simple. However, computer programs are needed for unresolvable mixtures, for example, when two contributors cannot be distinguished as major and minor. In mixture calculations, the concepts of restrict­ed and unrestricted come into play. In a restricted calculation, the relative peak heights at each locus are taken into account when pairing the alleles for the calculation. In an unrestricted calculation, all of the possible combinations of the alleles are deemed possible and are thus used in the calculation. The likelihood ratio calculation method in mixture DNA profiles where the peak heights and areas are not taken into account is called the unrestricted combinatorial method (Evett et al., 1991; Weir et al., 1997). This method examines all possible sets of genotypes consistent with the alternative hypotheses (prosecution and defense) and utilizes uniform assumptions such as the number of contributors across the loci. This method does not account for the possibility of ­drop-­out. There is a significant risk that the LR will be significantly ­nonconservative (Kelly et al., 2012). Better statistical models are now becoming available. These include ­semi-­continuous models like LoComatioN (Gill et al., 2007) or fully continuous models like TrueAllele (Perlin et al., 2011).



98



Introduction to forensic DNA evidence for criminal justice professionals



5.4 Mixture interpretation steps These procedures are modified from Clayton et al.,1998 and the SWGDAM guidelines (2010). Step 1: Identify the presence of a mixture—If more than two alleles appear at a locus, the presence of a mixture may be inferred. However, extra peaks may be present because of stutters and other artifacts. Allele symmetry may arise because of shared alleles and lead to masking; as a result the profile will appear unbalanced. Step 2: Designate allelic peaks—The peaks are designated as true alleles or as artifacts such as stutter, dye blobs, ­pull-­up, and so on. Step 3: Identify the number of contributors in the mixture— The number of alleles per locus, circumstances of the case, and possibility of related contributors factor into the determination of number of contributors. Step 4: Estimate the mixture proportion , i.e., estimate the ratios of the individuals contributing to the mixture—It may be possible to separate the mixture into major and minor components. In analyzing a mixture, the ratio or proportion of each contributor is approximately preserved throughout the mixture at each locus. Step 5: Consider all possible genotype combinations—All combinations of the unrestricted combinatorial list of genotypes are considered in relation to the mixture proportion and the heterozygote balance across all loci. Step 6: Compare reference samples—It is important that the previous steps take place without considering the reference samples to demonstrably avoid the possibility of bias. It may be necessary to consider different propositions at various stages of the analysis. Ultimately the court will decide those that are relevant for consideration. The prosecution and the defense both seek to maximize their respective probabilities of the evidence profile. There is no reason not to evaluate multiple pairs of propositions (Buckleton, 2005). The DNA result itself may indicate that different explanations are possible. One common misconception is that the numbers of contributors under the prosecution and defense hypotheses should be the same but there is no reason for this to be so. Both parties should confer with their respective forensic scientists to establish their hypotheses. The smallest numbers of unknown contributors are usually proposed in order to explain the evidence and maximize the desired likelihood. However, it is sometimes wise to denote different options for different numbers of contributors so that a court delay for calculations is not required.



Chapter five:  Partial profiles, low levels, and mixtures



99



5.5 Low template mixtures Stringent criteria should apply to the interpretation of mixtures and definition of peaks as true peaks. These criteria are difficult to meet in tests of touch DNA, as very small amounts of DNA may be recovered. In cases of mixtures, one of the contributors may be considered as low template DNA even though the total amount is 1 ng (say a 1:9 mixture). Thus issues of low template DNA must be considered. Incorporation of an assessment of the probability of allele ­drop-­in and ­drop-­out for low level DNA mixtures has been recommended. Another recommendation is that more empirical, quantitative data should be generated on the effects of low levels of DNA within mixtures. It is known that the stochastic nature of low template amplification renders any peak height threshold inaccurate below about 267 relative fluorescent units (RFUs). Therefore, caution has been urged for mixture interpretation when only trace amounts of DNA from one or more of the contributors are present (van Oorschot et al., 2010). Whenever d ­ rop-­out is a possibility, no meaningful exclusion probability can be calculated for the full profile. A random match probability (RMP) and a likelihood ratio (LR) approach may be extended to deal with situations where ­drop-­out is possible and no ­nonconcordant alleles are present. A n ­ onconcordant allele is present in the person of interest but not visualized in the electropherogram. The LR approach, but not the RMP, may be further extended to handle situations where ­nonconcordant alleles exist. In cases where ­drop-­out is invoked to sustain the prosecution case (not all the accused’s alleles are present in the crime DNA so there is ­nonconcordance of alleles), estimation of d ­ rop-­out probabilities cannot be avoided (Balding and Buckleton, 2009). For mixed DNA profiles of low template DNA exhibiting stochastic effects, the calculation of LR may proceed via either a binary, semicontinuous, or full continuous method (Kelly et al., 2012). The binary method treats alleles as present or absent. The ­semi-­continuous method assigns a probability to the events of d ­ rop-­out or no d ­ rop-­out but still treats alleles as present or absent. Fully continuous methods deal with the probabilities of stochastic events (like d ­ rop-­out) based on the heights of the peaks visualized at a locus. No modification of the binary method can deal with a ­nonconcordant allele in a comprehensive manner (Buckleton and Triggs, 2006). ­Drop-­out must not be a possibility for the unconstrained combinatorial method. Methods have been described for complex mixtures in which d ­ rop-­out is possible but there are no n ­ onconcordant alleles (Kelly et al., 2012).



100



Introduction to forensic DNA evidence for criminal justice professionals



5.6 Complex mixtures If a complex or indistinguishable mixture involves at least three individuals and no clear major contributor appears, L ­ R-­calculated yields limited evidential value. Sometimes a sample includes so many common alleles that few people are excluded, as in Case 4 from the author’s files. Case 4 An armed robbery by two masked men was committed at the home of a woman and her two young children. Accused B pleaded guilty and said Accused A was the main perpetrator. Touch and wearer DNA samples from the handles of a sports bag and the inner armpits and collar of a black suit jacket left in the backyard of the home were obtained by tape lifts. A statistical analysis found that Accused A was not excluded as a contributor to the DNA detected from the two items. LR statistics were performed considering two propositions for the DNA from the handles of the sports bag: (1) the DNA originated from Accused A, Accused B, and one other person chosen at random from the population or (2) the DNA originated from three other people chosen at random. The LR was determined to be 0.93. In other words, it was estimated to be 1.1 times less likely that the first proposition was true than if the second proposition was true. Statistics were also performed considering two propositions for the DNA from the suit jacket: (1) it originated from Accused A and two other people chosen at random or (2) it originated from three other people chosen at random. The LR was 0.4. In other words, it was estimated to be 2.5 times less likely that the first proposition was true than if the second proposition was true. There was more support for the proposition that the DNA from both the handles of the sports bag and the suit jacket came from three other people chosen at random from the population rather than for the proposition that the DNA came from Accused A and two other people (B and one unknown for the sports bag, or two unknowns for the jacket). Accused A was found not guilty after the jury deliberated less than 30 minutes. If a person of interest is included in a mixture and exhibits some common alleles, it may be harder to find two other individuals with the remaining (possibly rarer) alleles than it is to find three random people



Chapter five:  Partial profiles, low levels, and mixtures



101



who fit all the alleles in the profile. The inclusion of the person of interest is not surprising as almost everyone is included but “less included” than average. When a complex profile shows obvious stochastic effects, it is considered uninterpretable. The mixture classification scheme of the German Stain Commission (Schneider et al., 2009) denotes a mixture profile without major contributor(s) and with evidence for stochastic effects as “uninterpretable.” No method for interpretation currently exists for complex mixtures involving four or more persons, and thus such mixtures are not suitable for comparison to a person of interest. Results from simulation studies of a f­our-­person mixture (Paoletti et al., 2005) showed that 0.02% would show four or fewer alleles and that 76.35% would show six or fewer alleles using the U.S. 13-locus CODIS system. Consequently, more than 70% of f­our-­person mixtures would not be recognized as involving four persons based on allele counts. The studies showed a t­ hree-­person mixture would be incorrectly designated a ­t wo-­person mixture in a small percentage of cases. These studies also show the inherent problems of interpreting complex mixtures from three or more contributors.



References Balding, D. and Buckleton, J. 2009. Interpreting low template DNA profiles. Forensic Science International: Genetics, 4, 1–10. Bromwich, M. 2007. Final report of the independent investigation for the Houston Police Department Crime Laboratory and Property Room. http://www. hpdlabinvestigation.org/reports Buckleton, J. 2005. A framework for interpreting evidence. In Buckleton, J. et al., Eds., Forensic DNA Evidence Interpretation. Boca Raton, FL: CRC Press, pp. 27–63. Buckleton, J. and Triggs, C.M. 2006. Is the 2p rule always conservative? Forensic Science International, 159, 206–209. Butler, J. 2012. Advanced Topics in Forensic DNA Typing: Methodology. San Diego: Elsevier Academic Press. Butler, J. and Hill, C.R. 2010. Scientific issues with analysis of low amounts of DNA. Profiles in DNA, 13. http://www.promega.com/profiles/1301/1301_02.html Caddy, B., Taylor, G.R., and Linacre, A.M.T. 2008. A review of the science of low template DNA analysis. Home Office Forensic Regulation Unit. http:// police.homeoffice.gov.uk/publications/operational-­policing/ Caragine, T., Mikulasovich, R., Tamariz, J. et al. 2009. Validation of testing and interpretation protocols for low template DNA samples using AmpFLSTR Identifiler. Croatian Medical Journal, 50, 250–267. Clayton, T., Whitaker, J.P., Sparkes, R.L. et al. 1998. Analysis and interpretation of mixed forensic stains using DNA STR profiling. Forensic Science International, 91, 55–70. Crown Prosecution Service. 2010. Summary of national DNA database from the prosecution perspective. http://www.cps.gov.uk/legal



102



Introduction to forensic DNA evidence for criminal justice professionals



Evett, I.W., Buffery, C., Willot, G. et al. 1991. A guide to interpreting single locus profiles of DNA mixtures in forensic cases. Journal of Forensic Science Society, 31, 41–47. Fraser, J., Buckleton, J., and Gill, P. 2010. Review of DNA reporting practices by Victoria Police Forensic Services Division. http://www.vicpolicenews.com. au/images/stories/news/feature_story/victoria%20police%20forensic%20 services%20review%20%20report%20%20april%202010.pdf Gill, P. and Buckleton, J. 2010. A universal strategy to interpret DNA profiles that does not require a definition of low copy number. Forensic Science International: Genetics, 4, 221–227. Gill, P., Guiness, J., and Iveson, S. 2012. The interpretation of DNA evidence (including low template DNA). Forensic Science Regulator, Home Office Forensic Regulation Unit, Crown Copyright, July. http//homeoffice.gov.uk. Gill, P., Gusmao, L., Haned, H. et al. 2012. DNA Commission of the International Society of Forensic Genetics: Recommendations on the evaluation of STR typing results that may include ­drop-­out and/or ­drop-­in using probabilistic methods. Forensic Science International: Genetics, 6, 679–688. Gill, P., Brenner, C.H., Buckleton, J. et al. 2006. DNA Commission of the International Society of Forensic Genetics: Recommendations on the interpretation of mixtures. Forensic Science International, 160, 90–101. Gill, P., Kirkham, A., Curran, J. 2007. LoComatioN: A software tool for the analysis of low copy number DNA profiles. Forensic Science International, 166, 128–138. Gill, P., Whitaker, J., Flaxman, C. et al. 2000. An investigation of the rigor of interpretation rules for STRs derived from less than 100 pg of DNA. Forensic Science International, 112, 17–40. Hellmann, P. 2011. The H ­ ellmann-­ Zanetti report on the acquittal of Amanda Knox and Raffaele Sollecito. December 16. http://www.hellmannreport. wordpress.com Himmelreich, C. 2009. Germany’s phantom serial killer: A DNA blunder. Time (Berlin). Kelly, H., Bright, J., Curran, J. et al. 2012. The interpretation of low level DNA mixtures. Forensic Science International: Genetics, 6, 191–197. Mixture Interpretation Workshop. 2011. 22nd International Symposium on Human Identification, October 3. http://www.cstl.nist.gov/biotech/strbase/​training. htm Paoletti, D.R., Doom, T.E., Krane, C.M. et al. 2005. Empirical analysis of the STR profiles resulting from conceptual mixtures. Journal of Forensic Sciences, 50, 1361–1366. Perlin, M.W., Leger, M.M., Spencer, C.E. et al. 2011. Validating TrueAllele® DNA mixture interpretation. http://www.cybgen.com/documents/publications/ PLSSABD2010.pdf Schneider, P.M., Fimmers, R., Keil, W. et al. 2009. German Stain Commission: Recommendations for the interpretation of mixed stains. International Journal of Legal Medicine, 123, 1–5. SWGDAM (Scientific Working Group on DNA Analysis Methods). 2010. Interpretation guidelines for autosomal STR typing by forensic DNA testing laboratories. FBI website: http://www.fbi.gov/about-­us/lab/codis/ swgdam.pdf



Chapter five:  Partial profiles, low levels, and mixtures



103



Taberlet, P., Griffin, S., Goossens, B. et al. 1996. Reliable genotyping of samples with very low DNA quantities using PCR. Nucleic Acids Research, 24, 3189–3194. Thompson, W.C. 2009. Painting the target around the matching profile: The Texas sharpshooter fallacy in forensic DNA interpretation. Law, Probability, and Risk, 8, 257–276. Thompson, W.C. 2003. Review of DNA evidence in State of Texas v. Josiah Sutton. February 6. http://www.scientific.org/archive/Thompson%20Report.pdf van Oorshot, R., Ballantyne, K., and Mitchell, R. 2010. Forensic trace DNA: A review. Investigative Genetics, 1, 14. Weir, B.S., Triggs, C.M., Starling, L. et al. 1997. Interpreting DNA mixtures. Journal of Forensic Sciences, 42, 213–222. Word, C.J. 2011. Mixture interpretation: Why is it sometimes so hard? http://www.promega.com/resources/articles/profiles-­­in-­dna/2011



chapter six



Y-STR profiling A number of DNA techniques use different sequences from the autosomal STR DNA profiling methods described in the previous four chapters. This chapter will discuss Y chromosome short tandem repeat (­Y-­STR) profiling in humans. Chapter 7 covers mitochondrial DNA in humans and animals and more recent techniques such as familial DNA typing.



6.1 Introduction Continuing scientific discoveries leading to improvements and expanded applications of DNA are regularly reported in the (vast) literature and it may be a challenge for a legal practitioner to keep abreast of advances in the field. Details of the analysis of a forensic sample and consideration of the various DNA techniques employed should be described in a forensic science report as a guide to legal practitioners and other nonscientific reviewers. Legal practitioners should be aware of other discriminating techniques beside nuclear autosomal DNA profiling (discussed in the previous four chapters) that analyze human DNA for purposes of identification in criminal cases. These additional techniques utilize different primers and/­or sequences. Examples are the Minifiler STR that uses shorter tandem repeats, mitochondrial DNA profiling, and Y ­-­ STR profiling. Mitochondrial DNA and ­Y-­STR testing are genealogical techniques. The results from living individuals are often compared to historic populations or individuals. These tests are based on haplotypes (or complete sequence types) and are less discriminatory than autosomal STR profiling. The statistics for a Y chromosomal or mitochondrial DNA haplotype are different from those for STR autosomal DNA. The haplotypes must be treated mathematically as a single indivisible (atomic) trait. Thus, unlike traditional DNA methods that examine several traits that are approximately independent of each other, no multiplication of probabilities is possible with haplotypes. Therefore it is vital to have a sound fundamental understanding of atomic trait matching probabilities to make a reasonable assessment of the strength of identification evidence if these methods are used. ­Y-­STR profiling analyzes variations on the male (Y) chromosome in nuclear DNA. This technique can be used when autosomal (nuclear) DNA 105



106



Introduction to forensic DNA evidence for criminal justice professionals



typing is unsuccessful on a crime sample that contains male material and a possible donor needs to be determined. It is also particularly useful in cases where the DNA recovered from an item contains a mixture of male and female DNA. Since females possess only X chromosomes and no Y chromosome, this difference is exploited in order to target only the male DNA in a mixture containing male and female DNA. By contrast, males and females both inherit mitochondria from their mothers (see Chapter 7).



6.2 Benefits ­ -­STR typing may achieve profiles for male DNA (1) in samples containY ing low levels of male DNA and high background levels of female DNA, (2) in mixtures in which the female portion is present in overwhelming quantities compared to the male portion, (3) where there are multiple male contributors, and (4) in extended interval postcoital cervicovaginal samples. The following is a brief summary of situations in which Y ­ -­STR profiling may be worth considering (Roewer, 2009; Jobling and Gill, 2004): • Mixed stains in which the proportion of female DNA is higher than the male DNA present (frequently observed in vaginal swabs collected after sexual intercourse) • Cases of alleged sexual assault in which tests for seminal fluid or sperm are negative • Sexual assault cases in which the evidence is positive for semen, but no DNA foreign to the victim can be detected or potential male allele levels are below the threshold for autosomal STR detection • Sexual assault cases in which the evidence is ­ amylase-­ positive (­saliva-­presumptive) and a ­male-­­and-­female mixture is expected (e.g., traces of kisses or bites) • Cases with very old semen stains in which most sperm cells are suspected to be degraded and differential lysis is unsuccessful or risky • Sexual assault cases requiring screening of a large number of semen or other stains • Cases requiring determination of the number of male donors in a stain • Cases in which the evidence is expected to include cells of a male perpetrator (for example, underneath a female fingernail where male biological material may accumulate after a violent attack) • Cases requiring determination of the patrilinear relationship of a stain donor • Cases in which the stain donor’s population of origin must be inferred Another use is profiling in a case involving a very small number of male cells present at a rape scene because the rapist is azoospermic (lacks sperm)



Chapter six:  Y-STR profiling



107



or oligospermic (has a low sperm count). ­Y-­specific typing may be effective even if a vasectomized or azoospermic male leaves no sperm after coitus and the sample presents a 4000-fold excess of female DNA (Jobling and Gill, 2004). Autosomal STR analysis may not be useful if a sample contains an admixture of body fluids other than semen (saliva and saliva, saliva and vaginal secretion, fingernail scrapings revealing cells from a female victim and male perpetrator). It is not possible to use differential extraction to separate the male and female cells in such samples with current technology. The male component is often not detectable with the autosomal STR multiplex systems routinely used. Autosomal STR analysis may also fail with s­ emen-­containing samples in which very low copy numbers of sperm are present, or extremely fragile postcoital examples are taken after an extended interval (e.g., 48 hours after crime). Differential extraction of these samples may yield no profile from a male donor due to a combination of premature lysis of the cellular constituents of the sperm into the ­nonsperm fraction and sperm loss during the physical manipulations required for the isolation process. Therefore, Y ­ -­STR profiles that target only the male fraction eliminate the need for a differential extraction process and lessen the potential to lose the very small amounts of male DNA that may be present (­Mayntz-­Press et al., 2008). ­Y-­STR analysis presents several additional benefits in forensic casework. It allows the easy determination of the number of male contributors in a mixture. The profiles are “hemizygous” (one allele found at most loci) with the exception of a small number of multicopy loci). Multiple alleles at single copy loci clearly indicate the number of male contributors. Shortly after the characterization and evaluation of the first ­Y-­chromosomal STR polymorphism, its usefulness in crime casework was demonstrated when a mixed stain from a vaginal swab of a raped and murdered female victim was resolved by ­Y-­STR analysis and a falsely convicted male was excluded in 1992 (Roewer et al., 1992). Case 1 involved a request for retrial in Japan in 1998. The request was denied after 25-yearold vaginal swabs from two murder and rape cases were subjected to ­Y-­STR testing and matched the stains to each other and the defendant (Honda et al., 1999). Case 1 A retrial of a condemned criminal whose capital punishment had been suspended was requested by the Sapporo High Court in Japan. From 1972 to 1973, two successive rape and murder crimes and another rape occurred in Hokkaido in northern



108



Introduction to forensic DNA evidence for criminal justice professionals Japan. Two years later, a suspect was arrested. He confessed to the crimes during the trial and was sentenced to life imprisonment. However, he later insisted on his innocence. A second trial found that he was guilty and he was sentenced to death. The defense then appealed to the Supreme Court but the capital punishment was not overturned. After the judgment of the Supreme Court, the defense demanded a retrial on the ground that the judgment based on a confession was unsatisfactory—the original ABO blood grouping evidence was inconclusive. DNA was extracted from mixed seminal and vaginal secretions collected 25 years earlier from the two raped and murdered victims. Y ­ -­STR profiling was performed on the samples and four ­Y-­STR types were found identical to those of the accused. The High Court accepted the results and refused the retrial request in 1998.



­Y-­STR profiling may be useful in paternity testing, especially if the father is no longer alive. An example is the historically controversial case of Thomas Jefferson, the third president of the United States, who was thought to be the father of a child of Sally Hemings, one of his slaves. Jefferson had no sons so the descendants of his paternal uncle were subjected to Y ­ -­STR profiling. The Y ­ -­STR haplotypes were compared with male line descendants of the last son of Hemings. The results showed the same haplotype and supported the proposition that Jefferson could have been the father of Hemings’ child (Foster et al., 1998). In Case 2, the ­Y-­STR typing of a large population was used to eliminate suspects and then autosomal STR profiling was used to identify a serial rapist and murderer in Poland in the early 2000s (­Detlaff-­Kakol and Pawlowski, 2002). Case 2 A man committed at least 14 rapes in Poland since 1996 and murdered a 22-year-old woman in 2000. DNA profiles obtained from semen stains left at the crime scenes indicated that one male committed all the rapes. The Y chromosome haplotype obtained from the DNA in the semen stains was used to eliminate 421 suspects. One man exhibited a DNA profile identical at all Y chromosome STR loci analyzed and possessed common alleles in 9 of 10 autosomal loci. These findings strongly suggested



Chapter six:  Y-STR profiling



109



that the rapist and the man who exhibited the identical ­Y-­STR profile were closely related males. Analysis of reference DNA obtained from the man’s brother revealed an identical autosomal STR profile to those identified at the crime scenes. The U.S. case of A.B. Butler demonstrates the benefit of this method in excluding an individual convicted years earlier (Innocence Project). The accused was sentenced to 99 years and imprisoned in Texas in 1983 for kidnapping a woman from a parking lot and then raping her. The biological evidence was not tested until 1999 and autosomal DNA analysis did not yield conclusive results. Y ­ -­STR profiling had just been implemented in the New York Medical Examiner’s Office and the Texas evidence was then sent there. The results excluded Butler as the source of the semen from the rape kit. Butler was released in 2000 after serving 16 years in prison and was pardoned and compensated.



6.3 Theory The Y chromosome is paternally inherited and the profile is called a haplotype. These haplotypes are less diverse than the genotypes utilized in autosomal STR profiling containing equivalent numbers of markers. Patrilineal relatives (brothers, father, sons, paternal uncles) of a particular male will share a haplotype and this factor must be considered in any evidential analysis. The principal weakness of Y chromosome STR analysis is that even when a crime sample matches the profile of a suspect, patrilineal relatives of the suspect cannot be excluded as donors of the stain. Hence, in contrast to autosomal STRs, access to reference databases representing the variance and relatedness of haplotypes within local populations is crucial for interpreting Y ­ -­STR matches. The Y chromosome in humans is approximately 40 million base pairs long and contains just 78 genes. The ­sex-­determining region on the Y chromosome encodes a protein that triggers the development of the testes and, through an extended hormonal pathway, causes a developing fetus to become male. Most of the Y chromosome is n ­ on-recombining and passes unchanged from father to son except when mutations occur. This lack of recombination may be the reason why the Y chromosome reveals relatively few genes. Many Y ­ -­STR loci have been described in the literature (Butler, 2006). The basic repeats for most ­Y-­STR loci (as in autosomal STR profiling) are tetranucleotides (­four-­base pairs).



110



Introduction to forensic DNA evidence for criminal justice professionals



6.4 Statistics 6.4.1 Frequency estimates of ­Y-­STR haplotypes ­ -­STR profiling is very useful for exclusionary purposes because the Y result is unequivocal without the need to provide a statistical weight. If the Y ­ -­STR evidence is inclusionary, a statistical weight must be applied. The Y ­ -­STR loci are inherited from father to son as a single unit, virtually unchanged in each generation except for occasional mutations. Therefore, the haplotype of a man should be the same as his biological brothers, sons, and all other males along the paternal lineage. This haploidy and patrilineal inheritance complicate the interpretation of Y ­ -­STR haplotype matches because male relatives share identical ­Y-­STR profiles for several generations. Calculating statistics for Y ­ -­STR profiles is considerably different from developing statistics for autosomal DNA profiles. All ­Y-­STRs show linkage to the Y chromosome so that multiplying frequencies cannot be used to determine the frequency of a haplotype. Linkage and smaller effective population size contribute to ­population-­specific distributions that are affected by genetic drift and geographic differentiation. Population substructure effects have been shown to be more substantial for Y loci compared with observations for autosomal STR loci. Large databases of haplotypes must be maintained (typically by race or ethnic group) and the databases are then searched for haplotypes that match the haplotype of interest. Haplotype frequencies observed in or extrapolated from these databases often range between 1 in 1000 and 1 in 100,000, much lower than the “1 in 1 billion” typically cited in forensic reports for autosomal DNA profiling. The Y Chromosome Haplotype Reference Database (YHRD; www.yrhd.org) is an online facility designed to store Y chromosome haplotypes from global populations and replaces three separate database collections of European, Asian, and United States chromosomes (Willuweit and Roewer, 2007). As of February 2012, it contained over 100,000 haplotypes from more than 750 populations in 109 countries. The Y chromosome has a n ­onrandom distribution among global populations due to a practice known as patrilocality (the female moves to the male’s birthplace after marriage). Therefore, the priority for population sampling should not be sample size alone but should also include a good representation of the spectrum of p ­ opulation-­specific haplotypes. When sampled properly, even populations such as the Europeans, formerly regarded as sufficiently homogeneous for purposes of forensic genetics, appear genetically subdivided into distinct Y chromosomal clusters formed and maintained by recent demographic events (Roewer et al., 2005).



Chapter six:  Y-STR profiling



111



It is incumbent upon forensic scientists to assess the effects of population substructure and employ statistical approaches that address those effects for the relevant populations. The limitations of the specific database used and the meaning of any match found should be enunciated clearly in forensic reports.



6.4.2 Meaning of ­Y-­STR match A conservative statement for a Y ­ -­STR match report may be: The ­Y-­STR profile of the crime sample matches the Y ­ -­STR profile of the suspect (at xxx number of loci examined). Therefore, we cannot exclude the suspect as being the donor of the crime sample. In addition, we cannot exclude all patrilineal related male relatives and an unknown number of unrelated males as donors of the crime sample. Case 3 from the author’s files illustrates that DNA evidence can still be obtained even though initial attempts with autosomal DNA profiling were unfruitful due to a lack of spermatozoa. Case 3 The estranged partner of a woman was alleged to have burst into her home and raped her vaginally. He was also alleged to have put an axe handle in her vagina although it was never located. A single spermatozoon was determined from a high vaginal swab but no spermatozoa were found on any of the other swabs. No autosomal STR profiles were obtained from the medical swabs. A ­Y-­STR analysis was performed and a result obtained from the high vaginal swab (cellular fraction) and a labial swab, both with expected frequencies of 1 in 163 in the database (the haplotype was observed once in the database) that contained profiles of 1,079 individuals. The accused pleaded guilty at the beginning of the trial due to other issues so that the probity of the DNA evidence was not tested. Two current approaches are utilized to evaluate the probability of a coincidental match between two ­Y-­STR haplotypes if a frequency estimate is required. The most common method that has been used for many years with mitochondrial DNA profiling is the counting method (Gill et al.,



112



Introduction to forensic DNA evidence for criminal justice professionals



2001). The other is the haplotype surveying method (Roewer et al., 2005) which is a Bayesian approach that attempts to extract more information from the structure of Y ­ -­STR haplotype databases than does the counting method. An estimate of the frequency of the haplotype in a population is not possible by just calculating the number of matching profiles divided by the total number of profiles in a database. On many occasions, the specific haplotype has not been observed in the database (a null frequency), often due to the limited database size (see Case 3). A conservative bound is thus placed on the estimate to correct for possible sampling error. The confidence interval allows for a measure of the amount of confidence that may be placed on a value lying between two specified limits (the interval). One can calculate the upper bound of the confidence interval and this value can be used to convey, with a high degree of confidence, that the rarity of the evidence ­Y-­haplotype among unrelated individuals in a given population is less than the upper bound of the estimate. The assumption of a normal distribution may not apply for ­Y-­STR haplotype frequency estimates, but assuming normality will provide a conservative upper bound estimate. The Scientific Working Group on DNA Analysis Methods (SWGDAM) states that “… the use of the counting method that incorporates the ­upper-­bound estimate of the count proportion offers an appropriate and conservative statistical approach to evaluating the probative value of a match” (2009). The following calculations are based on SWGDAM guidelines. EXAMPLE 1 For a haplotype that is not observed in a database, the following formula is used to calculate the upper 95% confidence interval and serves as a correction for sampling uncertainty: 1 – (0.0.5) to the power of 1/n where n = size of the database. Assume that n = 2000, and there have been 0 observations previously in the database. Then 1 – (0.05) to the power of 1/2000 = 1 –0.9985032 = 0.0014967, or approximately 1 in 668. EXAMPLE 2 For a haplotype observed previously within a database, the calculation is: Upper bound = p + 1.96 √((p)(1−p))/n Lower bound = p – 1.96√(p)(1−p)/n



Chapter six:  Y-STR profiling



113



where p = x/­n, n = database size and x = number of observations of the haplotype in the database. Assume Mr. A has a haplotype that was observed five times previously (x = 5) in a database of n = 2000: Upper bound = 5/2000 + 1.96√((5/2000)(1−5/2000))/2000 = 0.0025 + 1.96(.00111663) = 0.0046886 (approximately 1 in 213)



A report should indicate that, “Mr. A could have contributed to the male source of the DNA detected. In addition, all male relatives on the paternal line and approximately 1 in 210 unrelated males cannot be excluded.” Appropriate caveats on the limitations of the database used should also be explained.



6.5 Number of male contributors to ­Y-­STR profile When more than one male contributes to a mixture, the individual male haplotypes may be difficult to distinguish unless they are present in different quantities. ­Y-­STR testing may still be useful for determining the number of male contributors in a mixture. Single copy ­Y-­STR loci usually produce a single amplicon in single source samples. Multiple peaks observed at such a locus may suggest two or more male contributors. While this is true in most cases, many regions of the Y chromosome are duplicated or even triplicated in some individuals and this can complicate mixture interpretation.



6.6 Determining mixture ratios It is possible to determine a major or minor contributor to a mixed ­Y-­STR profile. Unlike the autosomal STR assays, the peak heights across the range of molecular sizes do not remain relatively constant (maintain roughly the same peak height at different loci in the same dye) in Y ­ -­STR analysis. Thus, some of the smaller loci may disappear before the larger loci. For cases in which more than one haploid profile is detectable in a stain (for example, in a gang rape case), a likelihood calculation for varying numbers of known and unknown male contributors has been devised (Wolf et al., 2005). A prerequisite for such calculations is again the use of large ­Y-­STR haplotype population databases to retrieve frequencies of the haplotype profiles detected in the trace. However, two limitations of ­ Y-­ STR mixture analyses must be addressed: (1) the need to determine frequencies that apply for all possible



114



Introduction to forensic DNA evidence for criminal justice professionals



haplotypes that contribute to a mixture; and (2) u ­ ser-­friendly interpretation algorithms for Y ­ -­STR evidence profiles.



6.7 Combining statistics from autosomal and ­Y-­STR profiling No currently accepted method exists for combining autosomal STR and ­Y-­STR profiling. The different natures of underlying population structures infer that a combination of the information obtained from lineage genetic markers such as the Y chromosome (or mitochondrial DNA) with data resulting from meiotically recombining loci (autosomes) into a single likelihood ratio is inconsistent and should be avoided (Amorim, 2008). If calculations are performed for each type of profiling, details of both analyses should be conveyed in the forensic report. If inclusions drawn for both profile types support the hypothesis that a certain donor conveyed the DNA, the support cannot be quantified.



References Amorim, A. 2008. A cautionary note on the evaluation of genetic evidence from uniparentally transmitted markers. Forensic Science International: Genetics, 2, 376–378. Butler, J.M. 2006. Genetics and genomics of core short tandem repeat loci used in human identity testing. Journal of Forensic Sciences, 51, 253–265. ­Dettlaff-­Kakol, A. and Pawlowski, R. 2002. The first Polish DNA “manhunt”: An application of ­Y-­chromosome STRs. International Journal of Legal Medicine, 116, 289–291. Foster, E.A., Jobling, M.A., Taylor, P.G. et al. 1998. Jefferson fathered slave’s last child. Nature, 396, 27–28. Gill, P., Brenner, C., Brinkmann, B. et al. 2001. DNA commission of the International Society of Forensic Genetics: Recommendations on forensic analysis using ­Y-­chromosome STRs. International Journal of Legal Medicine, 114, 305–309. Honda, K., Roewer, L., and de Knijff, P. 1999. DNA typing from 25-­year-­old vaginal swabs using ­Y-­chromosomal STR polymorphisms in a retrial request case. Journal of Forensic Sciences, 44, 868–872. Innocence Project. www.innocenceproject.org Jobling, M.A. and Gill, P. 2004. Encoded evidence: DNA in forensic analysis. Nature Review: Genetics, 5, 742–751. ­Mayntz-­Press, K.A., Sims, L.M., Hall, A. et al. 2008. ­Y-­STR profiling in extended interval (≥ 3 days) postcoital cervicovaginal samples. Journal of Forensic Sciences, 53, 342–348. Roewer, L. 2009. ­Y-­chromosome STR typing in forensic casework. Forensic Science Medicine Pathology, 5, 77–84. Roewer, L., Arnemann, J., Spurr, N.K. et al. 1992. Simple repeat sequences on the human Y chromosome are equally polymorphic as their autosomal counterparts. Human Genetics, 89, 389–394.



Chapter six:  Y-STR profiling



115



Roewer, L. and Epplen, J.T. 1992. Rapid and sensitive typing of forensic stains using PCR amplification of polymorphic simple repeat sequences in case work. Forensic Science International, 53, 163–171. Roewer, L., Croucher, P.J.P., Willuweit, S. et al. 2005. International Forensic Y Chromosome User Group: Recent historical events in the European ­Y-­chromosomal STR haplotype distribution. Human Genetics, 116, 279–289. Roewer, L., Kayser, M., de Knijff, P. et al. 2000. A new method for the evaluation of matches in ­non-­recombining genomes: Application to ­Y-­chromosomal short tandem repeat (STR) haplotypes in European males. Forensic Science International, 114, 31–43. SWGDAM (Scientific Working Group on DNA Analysis Methods). 2009. Y-­ ­ chromosome short tandem repeat (­ Y-­ STR) interpretation guidelines. Forensic Science Communications, 11. Willuweit, S. and L. Roewer, L. 2007. ­Y-­chromosome haplotype reference database (YHRD) update. Forensic Science International: Genetics, 1, 83–87. Wolf, A., Caliebe, A., Junge, O. et al. 2005. Forensic interpretation of ­Y-­chromosomal DNA mixtures. Forensic Science International, 152, 209–213.



chapter seven



Other DNA techniques including mitochondrial DNA 7.1 Introduction This chapter describes DNA techniques other than autosomal DNA profiling (Chapters 1 to 5) and ­Y-­STR profiling (Chapter 6) used in criminal cases. Mitochondrial DNA is inherited maternally in the form of haplotypes and statistics are derived in a similar fashion to paternally inherited ­Y-­STR profiling. New and innovative techniques continue to be implemented in criminal cases along with combinations of mitochondrial and ­Y-­STR profiling. These techniques are often used as “last resorts” when autosomal DNA profiling is unsuccessful but are still very useful. A discussion below of the DNA analysis of bones from ancient and recently deceased humans and other species illustrates how DNA systems other than autosomal profiling can be used for identification purposes.



7.2 DNA analysis of bone DNA analysis of bones has been implemented forensically since work in 1991 demonstrated that DNA could be extracted from a corpse submerged underwater for 18 months and from bone marrow from the mummified corpse of an 11-year-old child (Hochmeister et al., 1991). An enduring mystery from the twentieth century was the fate of the last tsar of Russia, Nicholas Romanov II, and his family (Figure 7.1). The tsar abdicated during the Russian Revolution in 1917 and he, his wife, and five children were exiled to the city of Yekaterinburg. According to historical reports, all family members and their staff were executed by a firing squad in July 1918. A large mass grave was discovered in 1991 and DNA testing confirmed the identities of Nicholas, the tsarina, and three of their daughters in the grave (Gill et al., 1994). Nuclear DNA testing of five STR markers confirmed the sexes of the skeletons and established a familial relationship. Previous mitochondrial DNA testing confirmed an ancestral relationship through the maternal line between the Duke of Edinburgh 117



118



Introduction to forensic DNA evidence for criminal justice professionals



Figure 7.1  A 1914 photograph of the last Russian royal family. Seated from left to right are Grand Duchess Olga, Tsar Nicholas II, Grand Duchess Anastasia, Tsarevich Alexei, and Grand Duchess Tatiana. Standing from left to right are Grand Duchess Maria and Tsarina Alexandra. (Source: Harris and Ewing Collection, U.S. Library of Congress.)



(Prince Philip) and the tsarina and her three daughters that were found in the mass grave. Doubts persisted that these remains were in fact those of the Romanov family because the remains of the other two children (a boy and a girl) were missing. In 2007, human remains were discovered by amateur archaeologists in a small grave near the grave described above. A variety of DNA techniques linked fragments of bone and teeth from the small grave with bones from the large grave (Coble et al., 2009; Rogaev et al., 2009). The remains were badly damaged by fire and possibly sulfuric acid. Reference samples were provided by living relatives. The researchers were able to obtain complete mitochondrial DNA sequences from the charred bone fragments. Another link came through ­Y-­STR profiling that allowed a comparison of the Y chromosome markers (from the paternal lineage) of the purported male heir Tsarevich Alexei with the markers of the tsar and a number of living male descendants. Bloodstains from a shirt of the tsar found in storage at the Hermitage Museum in St. Petersburg* yielded a full autosomal STR profile and a ­Y-­STR profile that matched the putative remains found in the grave in Yekaterinburg. The shirt was obviously stored in an environment that



*



The tsar’s shirt was stored after he survived an assassination attempt in Japan in 1891.



Chapter seven:  Other DNA techniques including mitochondrial DNA



119



did not degrade the DNA from the blood, allowing a DNA profile to be developed more than 100 years after the blood was deposited. A debate over whether the remains of Anastasia or Maria were in the second grave could not be settled based on the study results (Coble et al., 2009). Over the years, many women have claimed to be Anastasia, including one named Anna Anderson. The results of mitochondrial DNA analysis of 20-year-old paraffin ­wax-­embedded samples and hair from Anderson ruled her out as a daughter of the tsar and tsarina (Gill et al., 1995). A number of mass graves dating from World War I were discovered recently in Fromelles in northern France. Some Australian and British soldiers killed in 1916 were buried behind what were then German lines. The graves were excavated in 2009 and 250 remains removed. LGC Forensics in England is performing ­Y-­STR analysis and mitochondrial DNA sequencing mainly on teeth but also bones; 75 bodies have been identified to date (Thomson, 2010). However, burnt, charred, and otherwise damaged bones present challenges in obtaining sufficient DNA for analysis. If DNA is exposed to fire or natural elements for any length of time, degradation can occur. A loss of signal may result from the presence of inhibitors and/­or the DNA is too fragmented to analyze. Thus careful optimization of all of the stages in the procedure of the analysis is mandatory. A study (Fondevilla, 2008) examined a charred femur from a major forest fire. ­Mini-­STR profiling and SNP (single nucleotide polymorphism) techniques were used to confirm identity by comparison with the alleged daughter of the male deceased. Mitochondrial DNA and ­Y-­STR profiling would not have been useful because mitochondrial DNA is carried maternally and Y ­ -­STRs are carried paternally. A more recent study has shown that the efficacy of obtaining DNA from burnt bones depends on the extent of burning (Schwark et al., 2011). Reliable DNA results could be obtained from well-preserved and ­semi-­burnt bones. The DNA of burnt black bones was highly degraded and often no nuclear DNA was left, leaving mitochondrial DNA as an option. Blue–­gray burnt bones yielded sporadic results and blue–­gray–­ white bones barely produced reliable results.



7.3 Mitochondrial DNA basics In addition to containing nuclear DNA, cell nuclei surround structures called mitochondria that essentially function like power plants by providing tools for cells to make energy. The mitochondria are about the size of bacteria and are scattered outside a cell nucleus. The mitochondrial genome is only 1/200,000 the size of the nuclear genome, occurs in many



120



Introduction to forensic DNA evidence for criminal justice professionals



(about 500 to 1,000) copies per cell, and is maternally inherited without recombination. Because mitochondrial DNA (mtDNA) is inherited maternally, brothers and sisters will have the same mtDNA as their mother, maternal aunts and uncles, and maternal grandmother. Mitochondrial results are much less discriminatory than those from autosomal DNA profiling. Mitochondrial DNA shares many of the theoretical disadvantages of ­Y-­STR profiling that were discussed in Chapter 6. It is n ­ onrecombining so that markers do not segregate independently, thus reducing diversity. It is uniparentally inherited through the mother so that members of the matrilinear line share the same haplotype. Mitochondrial DNA shows marked population substructure and presents the complication of heteroplasmy (see below). It has been accepted that ­Y-­STRs are easier to analyze than mitochondrial DNA. There are more haplotypes for ­Y-­STRs and they have larger population databases. Y ­ -­STR typing is performed at 12 or 17 loci in a single multiplex PCR assay, compared to mitochondrial sequence analysis across at least 610 nucleotides (and multiple strands) and often several amplifications with difficult samples. The major advantage of mitochondrial DNA is its multiple copy number per cell. This means that it has a greater probability of survival than nuclear DNA. Forensic applications include analysis of old, degraded and/­or damaged samples and samples such as hair shafts that contain low levels of nuclear DNA. Case 1 was the first U.S. criminal proceeding that introduced mitochondrial DNA profiling results from hair as trial evidence (Davis, 1998). Case 1 A Tennessee murder trial in 1996 convicted Paul Ware of the rape and murder of a 4-­year-­old girl. The defense claimed that the babysitter framed Ware, who was found drunk and asleep next to the body of the child. His semen was not found on the child. However, during the autopsy, a short red hair was found in the throat of the child and several red hairs were found in the bed at the crime scene. Mitochondrial DNA was extracted from two of the hairs— one from the throat of the victim and one from the bed where the offense was believed to have occurred. The mitochondrial DNA of the hairs was compared to and found to match Ware’s mitochondrial DNA. The haplotype had not been observed in an FBI database of 742 individuals.



Chapter seven:  Other DNA techniques including mitochondrial DNA



121



Case 2 (Innocence Project) further shows the advantages of mitochondrial DNA profiling on hair shafts. Evidence believed to have been destroyed was later located and was successful in clearing a wrongfully accused male. Case 2 In 1995, two men carrying pistols and wearing ski masks burst into an apartment in Oklahoma and bashed and robbed a young woman. She identified Sedrick Courtney as wearing a black ski mask (he lifted the mask before he bashed her). Black and green ski masks were discarded outside the victim’s apartment. Courtney was arrested despite having an alibi. The other robber was never found. Hairs were found in both ski masks and the roots were sent for autosomal DNA profiling; no results were generated. A forensic analyst stated that the hairs in the black mask could not be eliminated as having come from Courtney and a bleached red hair in the green mask was microscopically consistent with a bleached hair from Courtney. The prosecution stated the accused could have owned both masks and hairs from both masks could have come from him. In 2000 and 2007, Courtney requested DNA testing on the hairs but the police said the hairs had been destroyed. Courtney was paroled in 2011 and a lawyer for the Innocence Project learned that the police retained the hairs on a microscope slide. Mitochondrial DNA excluded Courtney from the hairs on the black and green masks and the charges against him were dismissed in July 2012. Mitochondrial DNA is defined by current convention into three regions: a coding region and hypervariable regions I and II. The DNA sequences of the hypervariable regions are usually analyzed as they contain the largest proportion of diversity in the mitochondrial genome. Heteroplasmy can cause different mitochondrial DNA sequences to be found in different tissues from a single individual, even along the length of a single hair shaft. Mutation, which distinguishes heteroplasmic types, is particularly common at some sites (hot spots). Shared heteroplasmy between two samples can actually increase the strength of evidence. This was shown in the confirming of the matrilineal relationship between the remains of Tsar Nicholas II and his brother, Georgii Romanov (Coble et al., 2009).



122



Introduction to forensic DNA evidence for criminal justice professionals



Case 3 is from a high profile 2004 California murder trial that involved a mitochondrial DNA match. Case 3 Scott Peterson was convicted in 2004 of murdering his wife (Laci) who was eight months pregnant with their child. The fetus was found washed up in San Francisco Bay in April 2003 and Laci’s torso was located the next day. The exact date and cause of death were unknown. Peterson reported his wife missing on Christmas Eve 2002. The only piece of forensic evidence was a six-inch black hair wrapped around a pair of pliers on Scott’s boat. His wife had never been on the boat. Mitochondrial DNA evidence from the hair was shown by the prosecution to be consistent with Laci’s mitochondrial DNA; her mother provided the reference sample. The defendant had a different mitochondrial haplotype. The haplotype of the hair from the pliers was expected once in every 112 Caucasians, from a database of 1,833 individuals. Defense lawyers challenged the evidence as unreliable. First, they declared that mitochondrial DNA was novel and not generally accepted in the scientific community. Second, the statistical probability in the case was insignificant and ambiguous and therefore incapable of helping the finders of fact in the dispute (Geragos, 2003). Alternatively, if the mitochondrial DNA met the standards, the defense protested that the careless actions of the police exposed the items to significant risks of alteration and contamination. The judge admitted the evidence (Girolami, 2003). Peterson has always maintained his innocence and is on death row; an appeal was filed in July 2012 (Elian, 2012).



7.4 Statistics in mitochondrial DNA analysis The first stage in assessing the value of a mitochondrial DNA profile is to consider the relevant population from whom the biological material may have originated. Similar to ­Y-­STR profiles, and because haplotypes from a database are compared, the relevant population depends on geographical factors and other circumstantial information (if any) known at the time. Could the donor of the material be a suspect in the murder of his business associate? Could a donor be the victim of a plane crash?



Chapter seven:  Other DNA techniques including mitochondrial DNA



123



The databases used may be different. In the absence of ­case-­specific information, international databases of mitochondrial DNA profiles that are available in print and electronically may be consulted to try to infer a donor’s ethnic appearance. The widely used EMPOP 9 database (www.empop.org) contains the profiles of about 4,500 individuals, mostly of European origin. The usual method of reporting the significance of matching profiles is to state the number of times (relative frequency) that this profile has been seen in the relevant database. For example, in the context of criminal proceedings in which Mr. X is the suspected donor of a hair with a certain mitochondrial DNA haplotype, an expert might state: In my opinion, the mitochondrial DNA sequence from Mr. X, which matches that of the crime stain, would be expected to be observed in fewer than 1 in 2,000 randomly chosen people in the European Caucasian population. A High Court of Australia appeal concerned the provision of two types of statistics in a murder trial (Ayturgrul v. The Queen, 2012, also discussed in Chapter 4). It was agreed in the trial that the appellant “could have been” the donor of a hair found on the deceased’s thumbnail. The statistics of a matching mitochondrial DNA profile were given as: (1) 1 in 1,600 people in the general population would be expected to share the same haplotype (frequency ratio); and (2) 99.9% of people would not be expected to have a haplotype matching the hair (exclusion percentage). The appellant asked to have the exclusion percentage figure judged inadmissible. He held that this percentage figure produced a subliminal or subconscious impact that invited the jury to approach the case with “percentages of guilt” and round the figure up to 100%. The prosecution expert used the counting method of observing the haplotype one time in a database of 4,839 individuals of various population groups. A defense expert was of the opinion that 1 in 1,000 people in the ­non-­Turkish population would have this haplotype and 1 in 50 people in the Turkish population would have the haplotype (the appellant was of Turkish descent). These appear to be very different statistics but they are dependent on their derivation. During the original trial, the method of deriving the statistics and the fact that mitochondrial DNA profiling was much less discriminatory than nuclear DNA profiling were discussed. The High Court dismissed the appeal as it did not consider that the appellant demonstrated that the probative value was outweighed by the danger of unfair prejudice.



124



Introduction to forensic DNA evidence for criminal justice professionals



7.5 Contamination A major concern in mitochondrial DNA analysis is contamination of a sample with extraneous DNA (Isenberg, 2005). Mitochondrial DNA analysis is a very sensitive technique and the presence of low level contamination is not uncommon. Contaminant DNA from any source can appear as a major or minor sample within a mixture or may overwhelm the target DNA completely. Gross or sporadic contamination can appear before an incident has occurred, in the interval between the incident and securing the scene, during the investigation of the scene, and/­or within the laboratory. Mixed mitochondrial DNA profiles in a bone sample indicate that contamination (extraneous DNA not inherent to the bone) has been introduced into the sample. A single bone or bone fragment belonging to a human should not have a mixed mitochondrial DNA profile that indicates contributors from at least two individuals. Case 4 from the author’s files was a homicide involving mixed mitochondrial DNA profiles in bone samples. Case 4 A man was reported missing in Melbourne, Australia. Bone fragments were found at a beach location about a two-hour drive away. It was alleged that the accused murdered the victim, burned his body in a drum, and disposed of the bones at the beach. One bone fragment failed to yield a nuclear DNA profile but gave a mixed mitochondrial DNA profile suggesting the presence of at least two mitochondrial DNA profiles. The prosecution expert stated that the bones were human due to the presence of mitochondrial DNA. Mitochondrial DNA from the bone powder was amplified and sequenced by another laboratory in the United States. Its sequence data also indicated a mixture of two or more mitochondrial DNA profiles and therefore the results were inconclusive. Because of the mixture profile of mitochondrial DNA and inability to assign a profile to the bone, the bone could not be determined to be human. If the bones were not originally cleaned sufficiently or a contaminant was introduced at the first examining laboratory, these errors will necessarily impact subsequent testing by another laboratory. However, if the cleaning techniques before the pulverization of the bones into powder met the standards, any touch DNA acquired through handling before introduction into the laboratory that remains on the surfaces of the



Chapter seven:  Other DNA techniques including mitochondrial DNA



125



bones has the possibility to be removed or considered. This is another reason to provide another analyzing laboratory with a whole item so that its contamination mitigation measures can be utilized. The mitochondrial DNA evidence was not admitted into trial. The accused was found guilty on other evidence. The ability of PCR (see Chapter 3) to amplify small amounts means that when ancient specimens contain little or no endogenous DNA, the DNA amplified may be derived partially or wholly from exogenous DNA contamination and be mistaken for endogenous (inherent) DNA. As an example, a report of dinosaur DNA sequences actually proved to be derived from human DNA contaminating the fossil (Malmstrom, 2005). All samples from 29 dog museum archaeological specimens contained human DNA often at levels exceeding authentic ancient dog DNA. Bones and teeth are normally the longest lasting physical evidence of human or animal presence and are also the most widely used samples for ancient DNA studies. However, they are readily contaminated (presumably through handling and washing) and difficult or impossible to decontaminate after such contamination. Sparse, damaged endogenous DNA is less likely to be amplified than modern contamination. Current techniques used to decontaminate specimens include the application of bleach, exposure to UV light, and grinding or shot blasting, and reflect a belief that contamination is concentrated in the outer surface of a material. Even when strict protocols are followed contaminants are frequently observed. Human DNA has been reported from cave bear samples, 500-year-old pig samples, and 109 of 168 relatively recent fox teeth. Several studies have reported significant numbers of human remains contaminated with multiple human sequences (Gilbert et al., 2005). Knowledge of the history of the sample handling prior to the analysis is thus critical. The important early Australian Mungo man study (Adcock et al., 2001) did not mention that the sample had been excavated in the 1970s and handled many times after that. Thus, it was likely to be contaminated, contained negligible organic preservation, and was considered too fragmentary to sex reliably because of very poor preservation practices. Without such information, it is very difficult to comment objectively on the reliability of results (Gilbert et al., 2005).



7.6 Mixture mitochondrial DNA profiles Unlike results from standard DNA and ­Y-­STR profiles, it is not currently possible to determine the number of donors or their respective contributions to a mixed sample using mitochondrial DNA. Some progress has



126



Introduction to forensic DNA evidence for criminal justice professionals



been made in separating mixed mitochondrial DNA profiles (Gilbert et al., 2005) but the techniques are not yet operational in forensic laboratories.



7.7 Familial DNA searching DNA databases are designed to provide law enforcement with investigative leads and rely on full or partial DNA matches (all alleles available) with crime scene samples. Familial DNA searching utilizes partial matches (not partial profiles) and further DNA interpretation is required. Close family relatives are used, for example a father and son who share one allele at each locus according to Mendelian inheritance (see Chapter 1). The United Kingdom pioneered the use of familial DNA searching. Case 5 was one of the first criminal matters utilizing this technique to solve the Valentine’s Day murder of Lynette White in 1988 in South Wales. Case 5 The victim was murdered in her bedroom. Investigators found more than 50 stab wounds, an almost severed head, and indications of male blood at the scene. Her flat was above a betting shop in Cardiff‘s “red light district.” Five local men, one of whom was her pimp, were arrested. The first trial did not finish due to the death of the judge. The second trial was held in 1990 and three of the five accused men were convicted of murder. However, the convictions were quashed on appeal in 1992. A cold case review was launched in 2000 and all the exhibits were sent to a different laboratory, Forensic Alliance (Exhibit A, 2004). The crime scene yielded 954 exhibits and a partial male DNA profile was obtained from a blood spot on a cellophane cigarette packet. It was postulated that one offender may have cut himself during the frenzied attack. The scientists went back to the original flat and found that it had been repainted. However they examined the skirting board below the original splashes of blood seen in the crime scene photographs. Three weeks of scraping back the paint revealed traces of the original bloodstains, and a full DNA profile was obtained from the blood that matched “cellophane man.” Eventually his blood was found in 10 places in the flat, on and around the body, and along the exit route. The DNA profile was placed on the DNA database but no match was found. However, the profile exhibited an uncommon allele, so the profile was searched on the South Wales



Chapter seven:  Other DNA techniques including mitochondrial DNA



127



police submissions database for a family type match. One profile that stood out was from a 14-year-old boy born after the murder. His uncle, Jeffrey Gafoor, gave a DNA sample to police and it matched cellophane man. He pleaded guilty and was convicted in 2003. When a routine search of a DNA database does not show that a crime profile matches any profile in the database, it is possible to conduct a search to identify potential relatives of the donor of a crime stain. This search is based on the number of shared genetic characteristics (alleles) and the rarity of the shared alleles in human populations. Unlike a search for a direct match, a familial search will allow for matching subsets of alleles at any given genetic marker as a basis for comparison. A familial search relies on mathematical modeling specific to the DNA database utilized. This modeling determines whether an observed similarity between two DNA profiles is more likely the result of kinship or mere chance (Myers et al., 2010). This analysis is more time consuming and labor intensive than traditional nuclear DNA testing. Not all jurisdictions have supported the technique and concern surrounds the value of familial DNA testing when balanced against privacy issues. Case 6 (Butler, 2012; Steinhauer, 2010) is a famous success story, although the trial has still not been held. At the time of the arrest of the accused, only two states (California and Colorado) had codified policies permitting familial searches. The “Grim Sleeper” case was the first use of an active familial search to solve a homicide in the United States. Case 6 Lonnie Franklin Jr. was arrested in July 2010 as a result of a familial DNA search in California. He is currently accused of the murders of 10 young women from 1985 to 2007. The women were murdered in Los Angeles and the cases were linked through firearms analysis and DNA. The perpetrator was called the “Grim Sleeper” because of the 13-year gap in detected crimes. A familial search of DNA database profiles in 2010 yielded one likely suspect based on the crime scene profiles from a DNA profile that was added to the database in 2009 after a felony weapons charge. Profiles from the Grim Sleeper crime scenes shared 1 allele at all 15 loci with the felon on the weapons charge. This meant that it was possible that the felon was a relative of the Grim Sleeper. They also shared the same ­Y-­STR profile.



128



Introduction to forensic DNA evidence for criminal justice professionals The Los Angeles police had a suspect, the father of the felon, and followed him. They assigned an undercover police officer to act as his waiter in a pizza restaurant and the waiter collected his discarded utensils and pizza leftovers so that a comparison DNA profile could be generated. The result was a full nuclear DNA match with the crime scene profiles. The accused pleaded not guilty.



The technique still remains controversial due to privacy concerns and technical pitfalls (Butler, 2012), but there is no doubt it can be invaluable when no investigative leads are generated.



7.8 Domestic animal hair Many households in every country keep domestic pets such as cats and dogs. These animals readily shed hair on the clothing and environment of humans (except for certain ­nonshedding breeds such as poodles). It is thus likely that this type of evidence can appear at crime scenes on the clothing of the victim and/­or perpetrator or on items like furniture. Today, the ability to profile DNA of animal hairs has increased the potential of this type of evidence. However, in general case work experience, this type of evidence is ignored or used as a last resort. It is worthwhile for legal professionals to recognize that animal hair may be used when other avenues have failed. Case 7 was the first criminal matter involving animal hair DNA. The hair came from Snowball, a white cat belonging to the parents of a murder suspect (­Menotti-­Raymond et al., 1997). Case 7 The body of 32-year-old Shirley Duguay was found in a shallow grave in a wooded area of Prince Edward Island in Canada in 1994, some eight months after she disappeared. A man’s jacket had been found 8 km from her house three weeks after she had gone missing. The leather jacket was covered in bloodstains matching the deceased and many white cat hairs were found on the lining. Douglas Beamish, the victim’s estranged boyfriend, lived with his parents who had a white cat called Snowball. The DNA profile of the cat hair on the jacket matched that of Snowball (­Menotti-­Raymond et al. 1997; Coyle, 2008). The case set a legal precedent allowing animal DNA to be admitted as evidence in criminal trials. Beamish was convicted of murder and sentenced to 15 years in prison.



Chapter seven:  Other DNA techniques including mitochondrial DNA



129



Cats have 18 pairs of autosomes and the X and Y sex chromosomes. The commercially available MeowPlex kit contains 11 STR markers (Butler et al., 2002). Dogs have 38 pairs of autosomes as well as sex chromosomes.



7.9 Other techniques A Minifiler STR kit uses STR markers that are reduced in size compared to standard STR kits used for routine DNA profiling (e.g., ProfilerPlus in Australia and the AmpFLSTR Identifiler in the U.S.). The reduced sized amplicons enable higher recovery of information from degraded DNA samples by improving amplification efficiency (Butler, 2012). DNA is now being applied to botany in cases where plant material from a scene may be linked to a suspect or victim. Microbial forensics is an emerging field that studies variations in bacteria and viruses. These new techniques must be shown to be valid and relevant before they are accepted as scientific evidence in courts of law.



References Adcock, G., Dennis, E., Easteal, S. et al. 2001. Mitochondrial DNA sequences in ancient Australians: Implications for modern human origins. Proceedings of the National Academy of Science of the USA, 98, 537–542. Ayturgrul v. The Queen. 2012. High Court of Australia, HCA 15. Respondent’s submissions to the High Court, S315, filed October 21, 2011. Butler, J.M. 2012. Advanced Topics in Forensic DNA Typing: Methodology. San Diego: Elsevier Academic Press. Butler, J.M., David, V.A., and M ­ enotti-­Raymond, M. 2002. MeowPlex: A new DNA test using tetranucleotide STR markers for the domestic cat. Profiles in DNA, 7–10. Coble, M.D., Loreille, O.M., Wadhams, M.J. et al. 2009. Mystery solved: The identification of the two missing Romanov children using DNA analysis. PloSONE, 4, March 11. Coyle, H.M., Ed. 2008. Nonhuman DNA Typing: Theory and Casework Applications. Boca Raton, FL: CRC Press. Davis, C.L. 1998. Mitochondrial DNA: State of Tennessee v. Paul Ware. Case Report. Profiles in DNA, GenePrint™. www.promega.com Elian, P. 2012. Scott Peterson appeals death sentence for 2002 murder of wife Laci Peterson. Huffington Post, July 6. Exhibit A. 2004. News from Forensic Alliance, 1. Fondevilla, M., Phillips, C., Naveran, N. et al. 2008. Case report: Identification of skeletal remains using short amplicon marker analysis of severely degraded DNA extracted from a decomposed and charred femur. Forensic Science International: Genetics, 2, 212–218. Geragos, G. (Attorney for Defendant Scott Peterson). 2003. Notice of motion and motion in limine to exclude mitochondrial DNA evidence. People of the State of California v. Scott Lee Peterson, Superior Court Case 1056770. October 20.



130



Introduction to forensic DNA evidence for criminal justice professionals



Gilbert, M.T.P., Rudbeck, L., Willerslev, E. et al. 2005. Biochemical and physical correlates of DNA contamination in archaeological human bones and teeth excavated at Matera, Italy. Journal of Archaeological Science, 785–793. Gill, P., Ivanov, P.L., Kimpton, C. et al. 1994. Identification of the remains of the Romanov family by DNA analysis. Nature Genetics, 6, 130–136. Gill, P., Kimpton, C., ­Aliston-­Greiner, R. et al. 1995. Establishing the identity of Anna Anderson Manahan. Nature Genetics, 9, 9–10. Girolami, A. (Superior Court Judge). 2003. Mitochondrial DNA evidentiary ruling. People of the State of California v. Scott Lee Peterson, Superior Court Case 1056770. November 18. Hochmeister, M.N., Budowle, B., Borer, U.V. et al. 1991. Typing of deoxyribo­nucleic acid (DNA) extracted from compact bone from human remains. Journal of Forensic Sciences, 36, 1649–1661. Innocence Project. http://www.innocenceproject.org Isenberg, A. 2005. Forensic mitochondrial DNA analysis. In Saferstein, J., Ed. Forensic Science Handbook, Vol. II. New York: Pearson Prentice Hall. Malmstrom, H., Storra, J., Holmgund, G. et al. 2005. Extensive human DNA contamination in extracts from ancient dog bones and teeth. Molecular Biology and Evolution, 22, 2040–2047. ­Menotti-­Raymond, M., David, V.A., and O’Brien, S.J. 1997. Pet cat hair implicates murder suspect. Nature, 386, 774. Myers, S.P., Timken, M.L., Piucci, G.A. et al. 2010. Searching for fi ­ rst-­degree familial relationships in California’s offender DNA database: Validation of a likelihood ­ratio-­based approach. Forensic Science International: Genetics, November. Rogaev, E., Grigorenko, A., Moliaka, Y. et al. 2009. Genomic identification in the historical case of the Nicholas II royal family. Proceedings of the National Academy of Sciences of the USA, 106. Schwark, T., Heinrich, A., ­Preusse-­Prange, A. et al. 2011. Reliable genetic identification of burnt human remains. Forensic Science International: Genetics, 5, 393–399. Steinhauer, J. 2010. ”Grim Sleeper” arrest fans debate on DNA use. New York Times, July 8. Thomson, J.A. 2010. The Fromelles War Graves Project: Identification of skeletal remains from World War I. Haploid DNA markers in Forensic Genetics Conference, Berlin. http://www.lgc.co.uk



chapter eight



Concerns and controversies This chapter will discuss situations in which DNA results may be compromised through the scientific processes that yielded them or through the interpretation of their meanings in the context of a case. Issues, such as contamination and error rates that affect the quality and subjectivity, and transfer that affects the reliability and validity of the results will be covered. A roll call of true but bizarre cases from England, Australia, the United States, and Europe will illustrate these points. At the end of this chapter, the obligations of the reporting scientist and the criminal justice professional involved in a case will be discussed.



8.1 Introduction Evidence can arise (1) through innocent means, (2) as a result of the crime, and (3) as a result of contamination or inadvertent transfer. The mechanism of transfer of a DNA profile is a consideration for every case reported. Two well known authors in the field emphasize the responsibility for a scientist to place the evidence in context and point out the limitations of interpretations (Gill and Buckleton, 2010). Limitations of the DNA results should be conveyed clearly in both the report and testimony of the forensic scientist. The legal practitioner should be alert to a few warning signs when examining a case involving DNA evidence. Cold cases are often ­re-­examined, but the exhibits may have been initially examined in conditions lacking the strict contamination mitigation measures used today. This is because transfer of minute quantities of DNA was not considered before the 1990s. Another warning sign is a single exhibit producing DNA results. The more DNA results of evidential value generated supporting the prosecution hypothesis, the less likely contamination, transcription, or other errors may have taken place. This is especially true if several body fluids or ­t wo-­way transfers are involved. Consider a case in which a DNA profile from blood found on the suspect’s clothing matches the victim and DNA from semen found on the victim matches the DNA of the suspect. Contamination would be less likely in this case than in a rape case in which DNA on a single medical swab from the victim matches the profile from the accused. 131



132



Introduction to forensic DNA evidence for criminal justice professionals



Finally, could the DNA have been deposited through legitimate contact based on the principles of trace evidence transfer? The method of deposition of the DNA should always be considered and various pathways proposed as a step in following the scientific method.



8.2 Quality issues Quality is considered the ability of a procedure or product to satisfy a need, be free of defects, and meet a set of requirements. Quality in a forensic science laboratory is controlled by a quality assurance system and should be strictly monitored, especially if the laboratory is accredited. A number of procedures may be followed if the quality of forensic evidence is suspected to be an issue in a criminal or civil case. These include audit trails of samples, validation studies of the test involved, peer reviews, internal and external audits, proficiency testing, and maintaining expertise. The processing of a DNA sample requires many steps and each step may present the potential for error. These steps include identifying the biological stain or material on the item, extracting the DNA, quantifying it, amplifying it, separating the components, and finally interpretation including statistical interpretation. Many cases are usually processed simultaneously, following each procedural step. Case 1 shows how contamination between two cases occurred in the extraction step of the DNA analysis. Case 1 Adam Scott, a 19-year-old male from Devon in England was accused of raping a woman in October 2011 in Manchester. He claimed he had never been to Manchester (Morris, 2012). Scott subsequently spent five months in jail on remand in custody after a database search allegedly found that his DNA profile matched that from semen found on a vulval swab of the victim. He was released in March 2012 after being found “the innocent victim of an avoidable contamination” (Rennison, 2012). Scott’s saliva from an unconnected earlier case preceded the pertinent medical swab sample taken from the alleged rape victim; it was in an earlier “batch” of samples through the extraction process. The DNA material from the victim’s medical swab was the sole evidence against the accused. DNA profiles from the seminal fractions of the two low vaginal swabs, high vaginal swabs, and a vulval swab matched the victim’s boyfriend. A second vulval swab produced a mixed profile from the victim’s boyfriend and an unknown male (17 of 20 alleles present). The unknown male DNA profile was



Chapter eight:  Concerns and controversies



133



loaded onto the national DNA database. The search yielded a partial DNA profile match of Scott’s DNA with a probability of one in one billion. The investigation revealed that a plastic sample holder tray that should have been discarded was mistakenly ­re-­used and loaded into equipment by a laboratory worker to undergo robotic extraction. Scott’s saliva sample had been processed earlier in the ­re-­used tray and in the same “well” (sample container slot) that was used later for the vulval medical swab from the alleged rape victim. It was determined that there was sufficient DNA present in the well from the earlier case to contaminate the later alleged rape sample. Guidelines in place in various jurisdictions around the world recommend that suspects should not be prosecuted on DNA evidence alone. These guidelines should serve as safeguards but they were not observed in Scott’s case. Continuity and audit trails may assist in querying the potential for error at each step. An audit trail of a sample should have a unique identifier that enables the sample to be followed at each stage of the analysis. This allows questions such as which scientist handled the sample on a particular date and location to be answered readily. The cases of O.J. Simpson (Case 2, Chapter 1) and the Omagh bombing (Case 1, Chapter 5) showed that poor collection procedures at crime scenes preceding exhibit arrival at a laboratory compromised both cases. Crime scene personnel today should follow the proper procedures for collecting exhibits and prevent the ready transfer of their own DNA to the scene and any exhibits they handle.



8.3 Relevant sample testing Every forensic report should incorporate the scientific method that includes a proposal of differing hypotheses and a testing rationale. A perusal of the case notes should elicit this information but if it cannot be deciphered, the legal professional should request clarification from the case-reporting scientist. Sometimes the relevant samples in a case are not tested. The bedding in the Frank Button case (Case 8, Chapter 2) was not tested initially. The defense later requested testing of the bedding and the result revealed semen with a DNA profile different from Button’s. The previously unsuccessful DNA result from medical swabs from the rape victim was also



134



Introduction to forensic DNA evidence for criminal justice professionals



r­ e-­analyzed and exhibited a profile matching that of a convicted rapist. Button was acquitted.



8.4 Contamination Contamination is an issue that may contribute to an error in a forensic result. It may result from a quality failure, for example, improper or careless handling introduces extraneous substances into a sample. Contaminants can be introduced at any stage of the testing process, from crime scene collection to the final result. Contamination of evidence has contributed to miscarriages of justice in many prominent cases. Cases 2 and 3 illustrate contamination of clothing exhibits. Case 2 The high profile murder investigation of the death of Jaidyn Leskie in Victoria, Australia involved laboratory contamination of evidence samples from different cases. The toddler’s body was found in a dam near Moe in 1998, some six months after he went missing. A trial jury acquitted the mother’s ex-partner and babysitter, Greg Domaszewicz, of the murder in 1998. A DNA profile was obtained in 2003 from the child’s clothing found in the dam; it matched the DNA profile obtained from a condom taken as evidence in a rape case. The police could find no connection between the rape victim and the murdered toddler. The inquest (Johnson, 2006) discovered that the child’s clothing was examined within days of the examination of the condom from the rape case by the same forensic scientist in 1998. The coroner found that contamination occurred in the laboratory although the exact pathway could not be determined. Another example of contamination between evidence samples occurred in the same Victoria laboratory during the cold case investigation of a double murder (Hadfield, 2011). Case 3 Margaret Tapp and her young daughter, Seana, were murdered in 1984 in Victoria. In 2008, a DNA database hit matched a profile from Seana’s nightwear with a reference sample from Russell Gesah. Detectives could find no link between Gesah and the Tapps and it was discovered that he was not in the state of Victoria



Chapter eight:  Concerns and controversies



135



at the time of the murders. An inquiry found that in 1999 an unrelated exhibit from Gesah was tested on the same day at the same laboratory as exhibits from the murder scene. Gesah was released from custody two weeks after his arrest; the murders are still unsolved. A report to the Victorian Parliament concerned the 2008 conviction of Farah Jama for the rape of an unconscious woman in a nightclub (Vincent, 2010; also see Chapters 2 and 4). The DNA evidence was contaminated in the examining rooms of a hospital. Justice Vincent noted in the report that the DNA evidence was perceived to appear so powerful by all involved in the case that none of the filters on which our criminal justice system depends to minimize the risk of a miscarriage of justice operated effectively at any stage of the case until a few weeks before Jama’s appeal. Vincent noted that no one appeared to be aware of the dangers of relying on statistical probabilities in the determination of guilt. The appeal court in the Meredith Kercher murder commented on contamination of exhibits collected at crime scenes, particularly low level DNA (Hellmann, 2011). An original statement was that it was not enough for the defense to claim that the DNA result was from contamination. The burden was on those claiming contamination to prove its origin. However, the appeal court held that the burden was proving that the result was obtained using a procedure that guaranteed the integrity of the item from the moment of collection to the moment of analysis. If no proof demonstrates that these precautions were followed, it is not necessary to also prove the specific source of the contamination (Hellmann, 2011). One way to prevent contamination is to ensure that reference and evidentiary samples are kept separate in time and space and that separation procedures are followed routinely in forensic laboratories. Prevention of contamination between evidentiary samples in a single case and even between cases, as seen from the above case studies, requires extra measures. No single specific test reveals contamination. However, audit trails may follow the path of an exhibit from collection to final analysis and indicate areas in which contamination may be a possibility. The laboratory environment in which DNA analyses are performed should be rigidly controlled. Leading U.K. forensic laboratories utilize requirements such as wearing disposable laboratory overalls, hair nets, face masks, and protective shoes on entry and disposing of the items on exit. Weekly monitoring of the background levels of DNA in laboratory rooms is also performed to ensure contamination mitigation. Poy and van Oorshot (2006) examined the levels of background DNA in a forensic laboratory. They studied laboratory and office areas at 195



136



Introduction to forensic DNA evidence for criminal justice professionals



sites and categorized them according to their contamination risk levels. Allelles were present at 52 of the sites (27%). Of 32 interpretable profiles, 28 matched staff members, mostly in the office area where no protective clothing was required; of the 4 unknown profiles 2 matched those in the criminal database. One profile recovered from a magnification lamp was believed to have come from a bulky, heavily stained sample examined 3 months prior to the study. The authors stressed the need for rigid contamination prevention. A further study was undertaken in Victoria to detect DNA on unused gloves from opened and unopened boxes in a forensic laboratory (Daniel and van Oorshot, 2011). The study revealed the presence of DNA on gloves from the closed boxes and on some gloves from opened boxes. It was recommended that only certified ­DNA-­free gloves be used in the laboratory and regular contamination monitoring of gloves be performed. An interesting study was performed in New South Wales to answer questions raised in two court cases. Allegations that police officers may have inadvertently transferred suspect DNA to drug balloons and firearms as a result of a search process were put forth (Beilby, 2006). The suggestion was made that transfers may have occurred through handling items belonging to the suspect (bag, clothing, other personal items), then handling other items while wearing the same gloves. The study found that secondary transfer to drug balloons was possible immediately after a search of a bag belonging to a volunteer, but did not occur if other objects were handled between the handling of the suspect’s items and the other evidence. Secondary transfer did not occur with firearms. The study recommended that police change gloves regularly during searches. Other avenues of contamination should be considered. Using collection and detection devices on multiple exhibits and at multiple scenes can introduce minor or gross contamination. Fingerprint brushes, for example, can transfer DNA between exhibits that could generate profiles. The brushes may also retain biological evidence for a considerable time (van Oorshot et al., 2010). DNA deposited on one item can thus be transferred to another. Case 4 from Germany confounded investigators and highlights the power of DNA evidence even when other evidence appears contradictory (Himmelreich, 2009). Case 4: The Phantom of Heilbronn Police linked 40 crime scenes incorrectly. One of these cases was the murder of a 22-year-old policewoman in the town of Heilbronn in southern Germany attributed to a female serial killer. DNA supposedly from the “phantom” (“woman



Chapter eight:  Concerns and controversies



137



without a face”) was found on a wide variety of items including a cookie, a heroin syringe, and a stolen car. In March 2009, it was revealed that all the cotton swabs used to collect samples from the 40 crime scenes may have been contaminated by the same female worker in a factory in Austria and no phantom ever existed. The phantom gained notoriety in 2007 after the murder of the policewoman. The police announced that they found the phantom’s DNA traces on several cold cases including a homicide from 1993. However, the types of cases committed (the phantom was both a brutal killer suspected of at least six murders and a common thief) were widely disparate. She had been involved in a car dealership burglary and a school ­break-­in. In both cases, her convicted accomplices denied her existence. The real phantom was discovered when officials trying to establish the identity of a burned corpse from fingerprints on an asylum application form found the form to contain the phantom’s DNA. The finding was considered impossible and officials repeated the analysis and found no DNA. They ultimately discovered that the cotton swabs used to collect DNA material were contaminated. Although they were sterilized for medical use, the sterilization process did not destroy DNA and some of the swabs contained enough cellular DNA to yield profiles. There had been a collective suspension of disbelief about the DNA evidence even among sophisticated detectives that trumped all other facts. As a measure to prevent cases like the phantom, the Scientific Working Group on DNA Analysis Methods (SWGDAM) in the U.S., the European Network of Forensic Science Institutes (ENFSI), and the Biologist Special Advisory Group (BSAG) in Australia have issued a position statement with specific recommendations for manufacturers and laboratories (Gill et al., 2010) to ensure that all materials used are D ­ NA-­free, rather than simply sterile. Police and crime investigators must invest greater effort in investigating and documenting how DNA samples arrived where found. Scientists must better understand the impacts of activities on the relative amounts of DNA from sources at a crime scene. In some instances, it is possible to derive the chain of events that led to the presence of a DNA sample at a crime scene, for example, prior visits to a scene or known use of an item. Awareness of these variables and their impacts on transfer events will



138



Introduction to forensic DNA evidence for criminal justice professionals



help weigh the likelihoods of proposed alternative scenarios (van Oorshot et al., 2010). Another case from Victoria showed that significant quantities of DNA are frequently (1) transferred from an exhibit to the inside of its packaging and (2) transferred from its area of initial deposit to other areas of the same exhibit and/­or to other exhibits within the same package (Goray et al., 2012). These findings highlight the need to deal with issues inherent in the collection and packaging of exhibits for forensic DNA analysis.



8.5 Interpretation issues Interpretation of DNA profiles is relatively straightforward when a sample is of sufficient quantity (all peaks above the stochastic threshold, see Chapter 5) and appears to be from a single source. It is even simpler when the DNA can be associated to a body fluid with some confidence. If confirmatory tests of these biological fluids are performed, the DNA may be associated more readily to a body material (although assumptions still need to be made and communicated). Problems in interpretation may occur when the DNA obtained is a mixture of contributors (two or more people), yields only a partial profile, is low level, or cannot be associated to a particular body fluid. Touch DNA requires careful interpretation. The murder of Meredith Kercher (Case 7, Chapter 2) involved touch DNA on a bra clasp that generated a minor DNA profile matching a profile of one of the accused. The result was a low level profile from a clasp collected 46 days after the crime. The delay in procuring the clasp introduced the potential for contamination at the scene. The appeals court also questioned why the knife in the case was believed to be evidential. The low level DNA on the blade of the knife found in a kitchen drawer of Sollecito’s flat could not be sourced to a body fluid. Furthermore, the victim and accused had access to each other’s apartments. Why was the possible presence of the DNA of the deceased on the blade considered evidential?



8.6 Error rates Forensic science commentators frequently ask the same question. Why is it necessary to have match probability data on DNA profiles to determine the rarity of a DNA profile but no data about the probability of a wrong result? The high numbers generated in DNA profiling of match probabilities appear to overwhelm all arguments. The probability of a wrong result is a different question. How often mistakes are made is a basic occurrence in science and is designated



Chapter eight:  Concerns and controversies



139



the error rate, but error rates in forensic disciplines have received little publicity. When DNA evidence was first introduced, a number of experts testified that false positives were impossible in forensic DNA testing. As we have seen in previous chapters, these claims are not true. Among the first 200 people exonerated by p ­ ost-­conviction DNA testing were Timothy Durham and Josiah Sutton (Innocence Project). Both were convicted due partly to DNA testing errors. In both cases, a combination of laboratory technical problems and careless or mistaken interpretation of the test results produced misleading evidence that helped send innocent men to prison for many years. There is little published data on error rates in forensic disciplines. Many forensic scientists argue that it is not possible to obtain an error rate in a specific discipline because there is no way to determine how often erroneous results are obtained. Some scientists such as fingerprint examiners claim that their discipline achieves a zero error rate. The assumption is that no two people have the same fingerprints. However, how likely is it that two people share a given number of fingerprint characteristics? No available data exist to determine that likelihood. One way to obtain error rates is to perform blind proficiency tests that mimic real cases, although objections have been raised to blind tests on the basis that proficiency testing can never reflect actual forensic casework. The only study to date examining error rates in forensic DNA analysis analyzed case data from 2008 to 2010 that involved over 200,000 DNA analyses (Kloosterman et al., 2012). The authors from the Netherlands Forensic Institute noted that it was impossible to compare their results with other studies despite the recommendation of the U.S. National Research Council to conduct research to study the sources of error of various forensic disciplines. The council described two types of errors in forensic DNA testing: Type 1 error: The DNA profile of a reference sample from a suspect is concluded incorrectly to match with the crime sample. Type 2 error: Wrongly reporting a ­non-­DNA match between two samples when in truth a match exists. The number of quality issue notifications constituted about 0.5% of the casework samples examined. The authors noted an important difference in impact between quality issues that have adverse outcomes on forensic investigations and failures that are recognized and corrected in the early stages of investigations. The authors also stated that research into the rates of type 1 and type 2 errors will provide only partial insight into the quality status of a laboratory. The actual frequency of these errors is low and does not



140



Introduction to forensic DNA evidence for criminal justice professionals



allow recognition of trends in error rates in the forensic community. They describe a system in place at the Netherlands Forensic Institute that relies on failure and potential failure classifications.



8.7 Overreliance on DNA technology DNA profiling is a test, and like all scientific testing is subject to quality control, even if automated systems such as robotics are used. Profiling is also performed and interpreted by fallible human beings. Transcription errors, handling mistakes, and other errors may contribute to incorrect results. Case 1 from England shows that contamination may occur even when automation and computers are used. A common axiom is that “computers do not make mistakes.” However, DNA analysis is still reliant on human beings at every step from isolation of the crime stain to final interpretation (see Chapter 3). Case 5 from the United States describes transcription error and/­or incorrect placing of samples in vials and involves a DNA database match that subsequently excluded the original accused. Case 5 A masked man in a blue hooded sweatshirt burst into a woman’s home in Las Vegas in 2001 and forced her to drive to an ATM for money. The man ran away when the woman’s husband spotted them (Mower and McMurdo, 2011). Police followed 18-year-old Dwayne Jackson and his cousin, Howard Grissom, who were riding bikes, because the police thought they could be the suspects. The police looked inside a car in the driveway of the suspects’ house and discovered a blue hooded sweatshirt with a ski mask in the pocket. Jackson’s DNA profile matched that on the sweatshirt; the DNA was the only evidence connecting him to the crime. Jackson pleaded guilty because the other charges of kidnapping and burglary that carried lengthy terms would be dropped if he did so. He was imprisoned for four years and released in 2006. In November 2010, the California Justice Department contacted Las Vegas police and informed them that someone else in the system matched the crime scene sample from the crime for which Jackson had been convicted. Grissom was convicted of an unrelated crime in southern California and was serving a 41-year jail term. His DNA profile matched the profile from the Nevada crime. It was discovered through a forensic review and



Chapter eight:  Concerns and controversies



141



r­ e-­analysis that a laboratory technician put Johnson’s sample into Grissom’s vial and vice versa in the original case. Jackson was awarded $1.5 million in a settlement in July 2011. The following bizarre case from England shows a consequence of transcription error. Case 6 Gareth Williams was a 31-year-old code breaker attached to MI6 at the time of his death in August 2010. His decomposing naked body was found inside a padlocked sports bag in the bathtub of his flat in Pimlico, London over a week after he was last seen (Davies, 2012). He had no injuries and no illicit or poisonous substances were found in his body. The keys to the padlock were beneath his naked body and the police determined he could not have locked himself inside the bag. A partial DNA profile belonging to a police scientist investigating the crime scene was discovered in February 2012. An inquest in April 2012 revealed that a typographical error of a forensic scientist in an email asking for a DNA check led police to believe there was foreign DNA on Williams’ body. Detectives wasted 18 months looking for a potential suspect who did not exist. The coroner determined that a still unknown party locked the victim inside his sports bag. The case is currently undergoing a forensic review.



8.8 Interpretation of DNA profiles: Objectivity and subjectivity Although DNA profiling is considered more objective than other forensic disciplines, scientist discretion still comes into play, particularly in interpreting partial DNA profiles or mixtures. Discretion is also a factor of crime scene selection, sample and item selection at the laboratory, tests performed, and subsequent interpretation and communication of the results. Crime scene samples are not pristine, and more often than not provide less quantity and quality than a scientist would desire. During the preparation of the appeal in the Jama case, the DNA statistics were recalculated by the forensic laboratory as a result of a review (Vincent, 2010). The chance probability of a match changed to 1 in 150 million for the Australian Caucasian population (from 1 in 800 billion



142



Introduction to forensic DNA evidence for criminal justice professionals



originally) and to 1 in 17 million for the Somali population (1 in 89 billion originally). Further calculations as a result of another review were likely to yield results somewhere between these values. How could this occur if the interpretation and statistical analysis of a DNA profile were objective? The controversy concerning interpretations of DNA profiles by the Victoria Police Forensic Services Department showed that interpreting DNA profiles involves opinion and subjectivity (Fraser, 2010). Designating a specific peak in a profile as true is subject to numerous factors, including its appearance and ability to meet certain criteria. A true peak must exceed a certain threshold value, have a certain height, be distinguishable from artifacts, and allow explanations of peak imbalances. Including particular peaks in mixture calculations is also a subjective process. One author describes subjectivity in interpretation of DNA profiles as akin to the ”Texas Sharpshooter Fallacy” or “painting the target around the arrow” (Thompson, 2009). The criteria for determining matches or inclusions for DNA profiles are purportedly shifted when a reference sample of a suspect becomes known according to the fallacy. Post hoc target shifting can distort the frequency and likelihood ratios, making matches appear more probative than they actually are. Epidemiologists named the “Texas Sharpshooter Fallacy” to describe the tendency to assign unwarranted significance to random data by viewing it post hoc in an unduly narrow context (Gawande, 1999). As an example, random cases of a disease may be interpreted to cluster in a particular population due to some particular cause or effect. The “painting the target around the arrow” description arose from the story of a legendary Texan who shot his firearm randomly into the side of a barn and then painted targets around the bullet holes. When the paint dried, he invited his neighbors to see what a great shot he was. They were impressed. They thought it extremely improbable that the shooter could have hit every target dead center unless he was indeed an extraordinary marksman and thus declared him to be the greatest sharpshooter in the state. In summary, the evidence of his accuracy was far less probative than it appeared. This fallacy acts as a type of confirmation bias because the human tendency is to interpret patterns in randomness where none exists. This type of fallacy may be applied to partial or incomplete DNA profiles (low levels and mixtures) subject to possible ­drop-­out and d ­ rop-­in. Such profiles present potential for subjectivity. Thompson (2009) notes that a key source of problems is the absence of formal standards for distinguishing inclusions from exclusions. An interesting study suggesting that DNA mixture interpretation may be subjective was conducted on an adjudicated criminal case (Dror and Hampikian, 2011). Seventeen North American expert DNA examiners



Chapter eight:  Concerns and controversies



143



from one laboratory in the jurisdiction of the original case were asked to interpret a mixture DNA profile and compare it to a suspect profile, without any contextual information given to examiners in the original trial. The experts came to conflicting conclusions about the inclusion or exclusion of the suspect. Most of the ­context-­free experts disagreed with the original pretrial conclusions, suggesting to the authors of the study that extraneous content influenced interpretation. The DNA evidence related to a gang rape. One of the assailants testified against the other suspects in return for a lesser sentence. However, those identified through the plea bargain denied involvement in the rape. The original result was that the DNA of the suspects identified could not be excluded from contributing to the mixture profile. The 17 examiners were asked to review the DNA mixture profile, particularly that of Suspect 3, originally determined as “cannot be excluded” in pretrial submissions. The evidence presented to the 17 examiners consisted of electropherograms relating to the sperm fraction from a vaginal swab and a reference sample of Suspect 3. One examiner reported that the suspect “cannot be excluded,” four examiners stated “inconclusive,” and 12 examiners said “exclude.” This suggested to the authors of the study an element of subjectivity in DNA interpretation. If the result was totally objective, all the examiners should have reached the same conclusion, especially since they all worked at one laboratory and used the same interpretation guidelines. It is desirable that all subjective judgments about an electropherogram be made without knowledge of the DNA profile of the person of interest who contributes to a mixed DNA profile (Kelly et. al, 2012). This should help prevent the raising of bias issues in court (Krane et al., 2008). For example, an examiner should determine whether ­drop-­out is possible at a locus before looking at a reference profile.



8.9 ­Retesting of samples On occasion, it may be desirable to r­ etest samples, for example, if the procedures used by the initial testing laboratory are in doubt because of contamination or some other error that occurred after an exhibit arrived at the laboratory. Sufficient uncontaminated sample left on the item examined or sufficient uncontaminated DNA extract must be available for retesting. During most testing protocols, sufficient DNA extract remains for ­re-­amplification by another laboratory. However, the sample remaining on an exhibit may not be adequate for a retest, especially if touch DNA is the suspect evidence. Some jurisdictions require retention of part of a sample (DNA extract) for further testing by the original laboratory. The volume remaining should be indicated in the case notes. The DNA extract should be stored



144



Introduction to forensic DNA evidence for criminal justice professionals



according to the laboratory protocol. Laboratories should maintain freezer systems specifically designed for DNA retention.



8.10 Adversarial system The Fraser team (2010) considered that different viewpoints are always present in adversarial systems, and where possible the differences are best resolved before trial. They noted that in Australia such pretrial meetings were not held or were not fruitful. In the United Kingdom, a pretrial review is a useful way to resolve disputes between defense and prosecution experts before any evidence is heard by a jury. The Fraser team felt the most appropriate responses to court challenge consisted of whether there was (1) active informed review of practices, (2) debriefing in problem cases, and (3) careful case preparation. The process of preparing for trial and reviewing evidence must involve all parties—the prosecution, the defense, the police, and the scientists.



8.11 Misconception about exact science The idea that science can be exact and objective in the sense of detachment from human judgment created many misconceptions about the nature of DNA profiling (Evett, 1996). One misconception is that there is an exact answer to the question of probability of a particular DNA profile given that it came from someone other than the defendant. However, the probability of an event is inevitably conditioned by the assumptions made. It is not possible in any situation to develop a probability without making at least one assumption. As illustrated in the Farah Jama case (Chapter 2) the chance match involved a number of different probabilities. Science is a complex subject involving many disciplines. Those with interest in the philosophy of science can refer to Aitken (1995) and Popper (1972).



8.12 Obligations Obligations are imposed on both the scientist who gives evidence at a criminal trial and on the representatives of the legal system who are responsible for conducting the trial. Justice Vincent (2010) reiterated this obligation related to DNA evidence in his review of the Farah Jama conviction. This obligation was enunciated also by a Royal Commission in Australia as early as 1984 in a case involving fiber and paint evidence (Shannon, 1984). A scientist should clearly and unambiguously describe the weight and substance that should be applied to scientific tests and specifically state the nature of any limitations.



Chapter eight:  Concerns and controversies



145



The responsibility that rests upon the criminal justice professional is to ask detailed and probing questions of the scientists in a case to elicit the required information and convey it to a jury. Chapter 9 provides some ideas for questions that should be asked of experts engaged by both prosecution and defense practitioners.



References Aitken, C.C.G. 1995. Statistics and the Evaluation of Evidence for Forensic Scientists. Chichester: John Wiley & Sons. Beilby, V. 2006. Court issues concerning secondary transfer of DNA. Australian and New Zealand Forensic Science Society’s 18th International Annual Symposium, Fremantle. Daniel, R. and van Oorshot, R.A. 2011. An investigation of the presence of DNA on unused laboratory gloves. Forensic Science International: Genetics, 3, e45–e46. Daniel, R. and van Oorshot, R.A. 2011. Environmental monitoring of background DNA within a forensic biology laboratory. International Association of Forensic Sciences 19th Conference, Madeira, September 12–17. Davies, C. 2012. Gareth Williams inquest hears of mystery DNA at crime scene. Guardian, April 24. Dror, I.E. and Hampikian, G. 2011. Subjectivity and bias in forensic DNA mixture interpretation. Science and Justice, 51, 204–208. Evett, I.W. 1996. Expert evidence and forensic misconceptions of the nature of exact science. Science and Justice, 36, 118–122. Fraser, J., Buckleton, J., and Gill, P. 2010. Review of DNA reporting practices by Victoria Police Forensic Services Division. http://www.vicpolicenews.com. au/images/stories/news/feature_story/victoria%20police%20forensic%20 services%20review%20%20report%20%20april%202010.pdf Gawande, A. 1999. The cancer cluster myth. New Yorker, February 8, 34–37. Gill, P. and, Buckleton, J. 2010. A universal strategy to interpret DNA profiles that does not require a definition of ­ low-­­ copy-­ number. Forensic Science International: Genetics, 4, 221–227. Gill, P., Rowlands, D., Tully, G.G. et al. 2010. Manufacturer contamination of disposable plastic ware and other reagents: An agreed position statement by ENFSI, SWGDAM, and BSAG. Forensic Science International: Genetics, 4, 269–270, 2010. Goray, M., van Oorshot, R.A., and Mitchell, J.R. 2012. DNA transfer within forensic exhibit packaging: Potential for DNA loss and relocation. Forensic Science International: Genetics, 6, 158–166. Hadfield, S. 2011. Man sues state over DNA bungle. Herald Sun, November 12. Hellmann, P. 2011. The H ­ ellmann-­Zanetti report on the acquittal of Amanda Knox and Raffaele Sollecito. http://hellmannreport.wordpress.com Himmelreich, C. 2009. Germany’s phantom serial killer: A DNA blunder. Time Magazine (available online). Innocence Project. www.innocenceproject.org Johnson, G. (State Coroner). 2006. Inquest into the death of Jaidyn Raymond Leskie. Coroner’s Case 007/98, Melbourne, Victoria.



146



Introduction to forensic DNA evidence for criminal justice professionals



Kelly, H., Bright, J., Curran, J. et al. 2012. The interpretation of low level DNA mixtures. Forensic Science International: Genetics, 6, 191–197. Kloosterman, A., Quak, A., Szerps, M. et al. 2012. Errors in forensic DNA casework: What types, how many, how serious? Paper presented at conference of European Association of Forensic Sciences. Krane, D.E., Ford, S., Gilder, J.R. et al. 2008. Sequential unmasking: A means of minimizing observer effects in forensic DNA interpretation. Journal of Forensic Sciences, 53, 1006–1007. Morris, S. 2012. Rape accused was victim of forensics error, regulator finds. Guardian, October 10. Mower, L. and McMurdo, J. 2011. Las Vegas police reveal DNA error put wrong man in prison. Las Vegas Review Journal, July 8. Popper, K. 1972. Conjectures and Refutations: The Growth of Scientific Knowledge. London: Routledge. Poy, A.L. and van Oorshot, R.A. 2006. Trace DNA presence, origin, and transfer within a forensic biology laboratory and its potential effect on casework. Journal of Forensic Identification, 56, 558–576. Rennison, A. 2012. Report into the circumstances of a complaint received from the Greater Manchester Police, F ­ SR-­R-618, September 17. Shannon, C.R. (Royal Commissioner). 1984. Royal Commission Report Concerning the Conviction of Edward Charles Splatt. Adelaide Government Printer. http:// nla.gov.au/anbd.­bib-­an3292187 Thompson, W.C. 2009. Painting the target around the matching profile: The Texas sharpshooter fallacy in forensic DNA interpretation. Law, Probability, and Risk, 8, 257–276. Thompson, W.C., Taroni, F., and Aitken, C.G.G. 2003. How the probability of a false positive affects the value of DNA evidence. Journal of Forensic Sciences, 48, 1–8. van Oorshot, R.A., Ballantyne, K., and Mitchell, R. 2010. Forensic trace DNA: A review. Investigative Genetics, 1, 14. Vincent, Justice F. 2010. Inquiry into the Circumstances that Led to the Conviction of Mr. Farah Abdulkadir Jama, Victorian Government Printer, Melbourne.



chapter nine



DNA pointers for criminal justice professionals 9.1 Introduction Reviewing a complex forensic report to determine what questions should be asked of the examiner and what areas must be challenged when discussing evidence in a criminal case is a daunting task. Ascertaining the quality of the examination may demand considerable time and effort. A full disclosure of laboratory records with a review by an independent expert is a very common tactic in English and American criminal trials. Sometimes experts are appointed by the court in European trials and act for the court and are not adversary witnesses. The complexity of DNA testing makes it difficult for a legal practitioner to evaluate the evidence without expert assistance. Looking behind the laboratory report to determine whether the underlying data support the conclusions should be the task of an expert witness. Experts may also assess whether alternative theories (inadvertent transfer, laboratory error, or sample contamination) of the evidence have been presented or considered. DNA profiles are complex and the propositions are also uncertain, so the legal practitioner may ask how the statistical analysis should be performed. An agreement regarding statistical analysis between the prosecution and defense is beneficial and the statistical model should be able to evaluate the differing positions. • The scientist is a facilitator of the discussion; strict boundaries must be observed. • The scientist must clearly define and separate the issues of relevance. He or she must eliminate confusion between the facts of the DNA profile and the circumstances by which the DNA was deposited at the crime scene. • The court then decides the relevant issues. The next sections summarize the many issues discussed in this book and may help a criminal justice professional understand and focus on DNA evidence encountered in a criminal case. 147



148



Introduction to forensic DNA evidence for criminal justice professionals



9.2 Advantages of DNA profiling • DNA profiling has a genetic basis and a firm scientific foundation. • DNA has the same composition throughout a body; DNA samples from blood and other body fluids can be compared. • DNA is relatively stable and that explains its utility in cold cases. • DNA profiling can be used on minute samples, even samples such as touch DNA that cannot be observed visually. • DNA profiling has very high discrimination potential and thus the power to exclude an individual is its fundamental advantage. • DNA profiling can provide high statistical value when a match is found; this lessens the possibility that another person has the same DNA profile.



9.3 Querying DNA evidence: Advice for the prosecution and the defense When do you call a forensic DNA prosecution expert to testify? If the prosecution calls an expert it is prudent to discuss the case with the expert to determine any strengths or weaknesses in the case. The prosecution’s failure to take this step may allow the defense to expose weaknesses of the results. Both parties should ensure that their experts are open and frank. Experts should be asked about possible quality failures, contamination, avenues of transfer, and inadvertent transfers. Audit trails and quality assurance procedures that prevent contamination may be readily apparent to an expert in the field. The details can also be explained privately to the lawyer in the case rather than in a courtroom. The lawyer may also be more prepared to examine the opposition’s forensic witness after discussions with his or her own expert. Issues discussed in this text, for example, hypothesis testing, subjectivity, and limitations, should always be considered by a presenting or challenging lawyer if forensic science evidence is crucial to a case. Cases involving mixtures of DNA require discussion between the prosecution and the defense to determine how many contributors may be represented in the mixture. The defense counsel and their experts should also discuss this issue. Usually the smallest number of contributors is proposed by the prosecution in order to explain the evidence and maximize likelihoods. If the number of contributors is vital to the case, it is wise to prepare and review the calculations to prevent delaying the court proceedings to make the calculations. Cellular material of an unspecified body origin, such as touch DNA, will need more consideration than DNA readily associated to blood, semen, or saliva. The modes of deposition must be postulated. The time



Chapter nine:  DNA pointers for criminal justice professionals



149



of deposition of DNA cannot be determined and that fact may or may not be critical to a case. If a haplotype is cited in a report, it refers to Y ­ -­STR or mitochondrial DNA profiling, not nuclear DNA profiling. Review the body of the report or any appendices for an explanation of the technique. The statistics in these types of profiles are much lower because of the inheritance of the DNA through the paternal or maternal line. ­Y-­STR and mitochondrial techniques are still very useful for exclusionary purposes. Technical problems may pose a major issue in a case. An expert may determine whether the controls used in the analysis accord as expected or present problems. What does a DNA profile (electropherogram) look like? Is it of good quality with all peaks above the stochastic threshold? Or is its quality poor, displaying many artifacts, overloading of DNA, inhibition, or ­drop-­out? Are the limitations of the techniques used described in the report and testimony? If not, why not? This book explains in detail the importance for forensic scientists to communicate both the limitations and advantages of DNA profiling. The collection methods utilized at crime scenes should be rigorous. Several cases in this book illustrate the pitfalls of poor collection practices. The appropriate personal protection equipment—gloves, face masks, hair nets, disposable overalls, and overshoes—should be used at all times. The gloves should be changed during collections of exhibits to prevent the ready transfer of DNA discussed previously in this book. The chain of custody must be complete and conform to anti-­ contamination measures in place. Evidentiary items for submission to DNA analysis should be bagged individually to prevent the transfer of DNA between exhibits within a bag and transfer of DNA from one part of an exhibit to the other (see Chapter 8). Revisiting a scene to collect exhibits may be unavoidable, but the investigator should assess scene security between visits.



9.4 Warning signs A number of warning signs may alert a legal practitioner when examining a case involving DNA evidence. Is one DNA result the only evidence in the case? Guidelines are in place around the world to ensure that no prosecution should occur if the sole evidence in a case comes from DNA. The Farah Jama case in Australia (Case 4, Chapter 2) and the Adam Scott case in England (Case 1, Chapter 8) show the consequences of unquestioned reliance on DNA. These individuals were imprisoned when in fact no offenses occurred. The DNA results of every case must be considered very carefully. A legal practitioner



150



Introduction to forensic DNA evidence for criminal justice professionals



should consider the DNA result, how the DNA was deposited, and how the DNA result fits in with the other circumstantial evidence in the case. The more DNA results of evidential value, the less likely contamination may have taken place. This is especially true if different body fluids or t­ wo-­way transfers are involved. Consider a case in which a DNA profile from blood found on the suspect’s clothing matches the victim’s DNA profile and DNA from semen matching the suspect’s DNA is found on the victim. Contamination would be considered a much smaller possibility in this case than it would be in a rape case involving a single medical swab from a victim containing semen matching the accused. Finally, could the DNA have been deposited through legitimate contact, using the principles of trace evidence transfer? Is the DNA relevant? If the accused and the victim cohabit, DNA transfers would not be surprising. What is more pertinent is the ability to associate DNA with a body fluid such as blood or semen. Cold cases are often ­re-­examined, but the exhibits may have been initially examined in conditions lacking the strict contamination mitigation measures used today. The reason is that transfers of minute quantities of DNA were not considered before the 1990s. The probative values of partial profiles, mixtures, and low level samples are particularly difficult to interpret and understand. Some reports do not state that a sample contained low level DNA. Low statistical values may or may not be an indication of low level DNA. Ask your expert whether a DNA profile is suboptimal. Could the DNA have been degraded or inhibited so that only a partial profile was obtained? Who are the contributors other than a suspect in a mixture of DNA? Were all necessary reference samples obtained so that they could be eliminated from the profile? Is masking an issue? Could a major contributor be separated out or were statistical packages used to deconvolute the contributors?



9.5 Was all evidence tested? There may be, for some reason, evidence that may be relevant or pertinent to the case that was not tested. Evidence may not have been collected from a crime scene or victim, or it may have been collected but not tested. The Frank Button case from Australia (Case 8, Chapter 2) provides an example; again the defendant served time in prison until another individual was proved to have committed the crime. A decision may be made not to test an item because of financial factors, time, sufficient evidence already obtained, limited resources, or direction from the submitting agency. Also a decision may be made to test only one stain. One semen stain on a sheet may be tested even if many were observed on the item.



Chapter nine:  DNA pointers for criminal justice professionals



151



9.6 Pretrial review A pretrial review should include all parties—prosecution, scientists, police, and defense team. Differing viewpoints are best resolved before trial. It is also beneficial if the parties agree to certain matters relating to DNA interpretation (e.g., numbers of contributors to a mixed DNA profile) before trial. The defense may request that its own expert ­re-­analyze the samples. If insufficient DNA extract remains from the sample in the original laboratory, it may be possible to examine the original exhibit and attempt to obtain additional DNA. However, all extract is often consumed in the original testing, especially in testing touch DNA. Debriefings of problem cases should be conducted by the prosecution, the defense, and the forensic laboratories.



9.7 Suggested cross-examination questions Before the trial (soon enough to allow an expert sufficient time to review the data), the legal professional should request full disclosure of laboratory records including case notes and methods used. Suggested questions are categorized and listed below.



9.7.1 General • Were the collection policies and practices at the crime scene or medical examination optimal in the analysis of this case? • Was a rationale for testing explained in the notes and or statement? If not, what was the rationale? • Was the scientific method used (and what is that?) • Have alternative hypotheses been considered? What are they? • Why is DNA profiling so powerful? (It has a high discrimination power and the power to exclude.) • Have the meanings of the scientific terms used been properly explained? • Was an i­ mpact-­based priority testing system used? • What quality assurance procedures were in place? • Is the examiner aware of observer and/­or context effects? • Does the examiner know the error rates of the tests? Can he or she explain this concept? • How have the statistics quoted in the report been determined? • Is there a possibility of transfer (primary, secondary, or higher)? • Did the positive and negative controls perform as expected? • Does the laboratory have databases for investigating contamination events including elimination databases for consumable suppliers (where possible), police officers attending crime scenes, crime scene



152



Introduction to forensic DNA evidence for criminal justice professionals



operators, and laboratory staff (scientific and administrative)? What are they specifically? • Were there issues with the technical review? • Were there issues with the administrative review?



9.7.2 Single source DNA profiles associated with blood, semen, or saliva • • • • • • • • •



Can the DNA profile be related to a specific body fluid? If so, how? What reference profiles were used and how were they obtained? Were all appropriate reference samples taken and profiled? Was the evidence profile interpreted and designated as single source before comparison with reference DNA profiles? How can we be sure that contamination was prevented? What are the limitations of the results? Have the appropriate population databases been used? Was extra scrutiny applied if there is only one DNA result from many items tested? If there is an inclusion, or a match, what is the statistic and what does it mean?



9.7.3 Difficult DNA profiles (partial, low level, mixture, unspecified origin) • Is this a partial DNA profile and why? • Does this DNA profile exhibit degradation or inhibition and why? Was the sample ­re-­amplified to obtain a better result? If not, why not? • Are any of the samples mixtures from two or more individuals? • Can the mixtures be separated into major and minor contributors, and if so why and how? • What are the possible methods of transfer of the DNA? • Can the DNA detected be related to a particular time? • Can the DNA be related to a particular body matter? If so, how? • Do any profiles exhibit low level DNA and require extra scrutiny? • Are any of the peaks in the profile below the stochastic threshold? How has the witness dealt with this? Would the witness say this profile is suboptimal? • If the profile is low level, what extra precautions were taken, if any? • How was the final profile derived—through consensus profiles or the statistical model? What was the rationale?



Chapter nine:  DNA pointers for criminal justice professionals



153



9.7.4 Expert witness • Does the witness have an appreciation of DNA interpretation practices internationally? (References cited in this book may be useful for preparing questions.) • Does the witness participate in a continuing education program?



9.8 Discovery requests This section summarizes the documents required for disclosure of scientific materials pertaining to DNA testing. The summary is applicable to all DNA testing that has been performed, is in progress, or planned for the future. Figure 3.6 in Chapter 3 lists the documentation that should be included in every case file. Case file—Documents listed in Figure 3.6 including all records generated by the testing laboratory and copies of all photographs should be compiled and available for discovery. ­Case-­related correspondence between investigating police, other officials, and staff members should be included. Laboratory protocols—Copies of all standard operating protocols used in connection with laboratory testing should include explanations of the method of DNA analysis, the statistical interpretation details, pertinent validation studies, population databases, and allele frequencies. Chain of custody—All records that document the handling of the evidence from the initial point of collection to current disposition should be retained for discovery. The records should indicate how the materials were stored (temperatures and types of containers), the amount of evidence consumed in testing, the amount of material remaining, and where and how the remaining evidence is stored. Software—A list of all commercial software programs used in the DNA testing should include name, manufacturer, and version used. Data files—Data describing extraction, quantification, amplification, separation, and analysis should be maintained in case files. The data should indicate the dates when steps were performed and names of personnel performing tests and checking results. STR frequency tables—Copies of allelic frequency tables used in making statistical estimates should be retained. If the testing laboratory relied upon published data, this requirement may be satisfied by specific references to sources. Unintended transfer or sample contamination—Records maintained by the testing laboratory should document instances of unintended transfer or sample contamination and describe all corrective measures taken. Accreditation—Is the laboratory accredited? Has this accreditation ever been suspended? If so, for what reason?



Appendix A: Glossary of terms used in reports and testimony



Allele:  Alternative at a site on a DNA molecule; one alternative is inherited from each parent; variation at a given locus on a chromosome; number of short tandem repeats at a locus on a chromosome. Allele ­drop-­in:  Contamination from a source not associated with a crime stain and manifested as one or two alleles. Allele ­drop-­out:  Low level of DNA insufficiently amplified to yield a detectable signal. Amelogenin:  Part of DNA strand used to characterize the sex of an individual contributing DNA. Amplification:  Process by which the number of copies of specific DNA sequences is increased via sequential copying. Analytical threshold:  Minimum height requirement at and above which detected peaks can be distinguished reliably from background noise on an electropherogram. Artifact:  Result appearing in DNA profile (electropherogram) that arises from the process and is not intrinsic to the DNA tested; also see stutter and ­pull-­up. Autosome:  Chromosome not involved in sex determination; humans have 22 pairs.



155



156



Appendix A: Glossary of terms used in reports and testimony



Base pair:  DNA is formed from four chemical bases; a base pair (bp) consists of a base in one strand of the double helix and its complementary base on the other strand. Buccal swab:  Mouth swab used to obtain DNA samples. Chromosome:  Discrete unit of the genetic material carrying genes; chromosomes are arranged into structures that can be visualized during cell division. ­Co-­ancestry coefficient:  Allowance for possible shared ancestry within a population; designated between 0 and 1; higher values correspond to greater shared ancestry. Confirmatory test:  Test used to confirm the presence of a specific biological material such as blood or semen. CODIS:  Combined DNA Index System database for DNA profiles used in the United States. Conservative:  (1) Assignment for weight of evidence that is believed to favor the defense. (2) When the evidence is very powerful in one direction, assigning a weight below the level of belief in that direction. (3) Lack of conservativeness often results when the assumptions underpinning a statistical model are seriously violated. Contamination:  Extraneous DNA from a source not associated with a crime stain, for example plasticware can be contaminated at a manufacturing source. Crime scene sample:  Sample taken from a crime scene or body by crime scene investigators or medical personnel. Degradation:  Breakdown of DNA strands through age, environment, or chemical insult resulting in a greater loss of longer fragments. Deoxyribonucleic acid (DNA):  Chemical compound found in all nucleated cells of a body. It codes for characteristics in humans. Electropherogram:  Output of electrophoresis instrument that displays DNA profiles as peaks on a graph. Electrophoresis:  Method of separating molecules based on size and charge, used to separate DNA fragments of varying length by application of an electric current. Exclusion:  Exclusion of a contributor to stain. (1) Decision by an expert that a certain reference DNA profile does not represent a contributor to the stain. (2) Situation in which a reference profile is “excluded.” (3) Exclusion from a stain at one or more loci. (4) Exclusion at a locus. (4) Pattern of assumed genotypes at a locus indicating that an allele seen in a certain reference DNA profile is not observed in a stain. Exclusion probability:  Probability of the exclusion of a randomly selected DNA profile.



Appendix A: Glossary of terms used in reports and testimony



157



Familial searching:  Process allowing potential relatives of offenders to be identified in a DNA database when the offender profile is not in the database. Frequency:  Rate at which an event occurs. For example, sample frequency of an allele is the number of occurrences of the allele in a population sample divided by the sample size: population frequency of a DNA profile is the (unknown) number of times that the profile occurs in the population divided by the population size. Gene:  Site on a DNA molecule; sequence of the inherited code for which there is a functional product; sequence of DNA base pairs. Genome:  Total genetic material of an organism contained in a full set of chromosomes. Genotype:  Characterization of alleles at a specific site; the designation of two alleles at a locus is a genotype. Haplotype:  Collec­tive genotype of a number of closely linked loci on a chromosome that are inherited together; in mitochondrial DNA, the sequence of the control regions that pass unchanged from a mother to her offspring; in Y-STR typing, a Y-STR sequence that is inherited unchanged from male to male. Heterozygote:  Two alleles of an individual at a particular site are different. Homozygote:  Two alleles of an individual at a particular site are the same. ­Intelligence-­led screen:  Screening conducted during major crime investigations applied to a DNA sample from a number of people who may have been associated with a crime or crime scene with the intent to eliminate them from the inquiry. Likelihood:  Conditional probability of an event; an event is considered an outcome corresponding to one of several conditions or hypotheses. Likelihood ratio (LR):  Ratio of two probabili­ties of the same event under different hypoth­eses. The numerator typically contains the prosecutor’s hypothesis and the denominator is the defense hypothesis. The LR is often expressed as the ratio between the likeli­hood that a given profile came from a particular individual and the likelihood that it came from a random unrelated person. Locus:  Area of DNA that is analyzed to generate a DNA profile. Loci is the plural. Low level or low template analysis: Analysis result from very small amounts of DNA that typically produces peaks below the stochastic level. Masking:  Result from analysis of a DNA mixture showing overlap of the same allele originating from different contributors. Mitochondrial DNA:  DNA obtained from the mitochondria of a cell.



158



Appendix A: Glossary of terms used in reports and testimony



Mitochondrion:  Subcellular unit within the cell that provides energy. Mitochondria is the plural. Mixture:  DNA typing result originating from two or more individuals. Mixture ratio:  Relative ratio of DNA con­tributions of multiple individuals to a mixed typing result based on quantitative peak height information. A mixture ratio may also be expressed as a percentage. Molecule:  Chemical substance consisting of atoms bound together in a specific structure. Multiplexing:  Method of amplifying multiple sites on a DNA molecule in a single reaction vessel. Nuclear DNA (nDNA):  DNA found in cell nucleus. Nuclear DNA testing consists of autosomal STR DNA typing and ­Y-­STR DNA typing. Nucleus:  Cell structure that contains most of the DNA. Nuclei is the plural. Partial profile: Result obtained from a sample deficient in quality or quantity; no full profile is produced. Polymerase chain reaction (PCR):  DNA amplification process by which one or more DNA regions are copied using a DNA polymerase enzyme to generate enough DNA for analysis. Presumptive test:  Screening procedure that indicates the possible presence of a specific body fluid. Primer:  Reagent containing synthesized DNA sequences that are complementary to a specific segment of the DNA on either side of the area of interest used in the PCR process. Each primer is labeled fluorescently so that the copied STR allele can be detected visually and its length measured Probability:  Rate of occurrence of an event in a repeatable experiment; expected frequency; number between 0 and 1 that reasonably reflects the belief that an event is true. Profile:  Numerical format showing STR alleles detected; one or more genotypes used for DNA comparison. ProfilerPlus:  Technique used in Australia to profile DNA samples. It analyzes 10 areas of a DNA molecule including the sex chromosome. Proposition:  Hypothesis of the defense or prosecution arguments used to formulate a likelihood ratio. ­Pull-­up:  Artifact resulting from fluorescent detection. Random match probability (RMP):  Probability of randomly selecting an unrelated individual from a population who has the same DNA profile as that of the questioned sample. Relative fluorescent unit (RFU):  Unit of measurement of peak height on an electropherogram resulting from detection of fluorescence intensity. Restricted combinatorial method:  Elaboration of the unrestricted method in which allelic intensity (peak height/­area) data are used to restrict the sets of genotypes that are considered plausible explanations.



Appendix A: Glossary of terms used in reports and testimony



159



Resolvable DNA mixture:  Mixture of DNA from two or more individuals detected on an item of evidence in which the ratios of major and minor con­tributors can be deduced due to the proportion of one versus the other. Sex chromosome:  Chromosome involved in determining the sex of an individual. Females possess two X chromosomes and males possess one X and one Y chromosome. Short tandem repeat (STR):  Type of locus used to generate DNA profiles through nuclear typing. Stochastic effect:  Imbalance in the amplification of two alleles by competition during PCR; one allele is preferentially amplified over the other. Stochastic effect, random:  Variations in alleles on repeat sampling with small amounts of DNA. Stochastic threshold: Minimum DNA quantity required to produce a reliable or optimum profile, assuming that d ­ rop-­out of a sister allele has not occurred at a particular locus above this threshold determined in specific laboratory studies. Stutter:  Allelic artifact caused by slippage of the polymerase enzyme which is four bases fewer than the allele that causes the stutter. ­Over-­stutter (one base pair longer than its associated allele) occasionally occurs. Stutters are always found in allelic positions and can compromise interpretation of minor contributions to mixtures. Touch DNA:  DNA from skin (epithelial cells) that is left behind when a person touch­es or otherwise comes into contact with an item or other person. Unrestricted combinatorial method:  Simple likelihood ratio method of evaluating mixture evidence. The method assumes a list of all alleles in the mixture and considers competing hypotheses that various known or unknown profiles are constituents of the mixture. It uses no data on allelic intensities, hence one set of genotypes whose allele sets are coincident with the mixture is considered to have as valid an explanation of the mixture as any other set. ­Y-­STR profiling:  DNA typing in which STRs on the Y (male) chromosome are analyzed.



References Gill, P., Brenner, C.H., Buckleton, J.S. et. al, 2006. DNA Commission of the International Society of Forensic Genetics: Recommendations on the interpretation of mixtures. Forensic Science International, 160, 90–101. ­Puch-­Solis, R., Roberts, P., Pope, S. et al. 2012. Practitioner Guide 2: Assessing the Probative Value of DNA Evidence. London: Royal Statistical Society, London.



Appendix B: Selected DNA issues and case examples



Chapter 1



Case 1 4



2



3



6 7



1 2 3 3 5 2 1 2 5 6 7



Description Notable cases solved by DNA Colin Pitchfork case; first murder solved by DNA profiling; first DNA-led screen; first person exonerated by DNA Kirk Bloodsworth case; first conviction overturned by DNA testing Ray Krone case; contrast of DNA with questionable science Damon Thibodeaux case; questionable assumptions about absence of evidence Robert Dewey case; stability of DNA Darren Paul Smith case; cold case solved through DNA database Analysis of DNA from maggots ingesting body Mass screening through Y-STR analysis Paul Ware case; mitochondrial DNA Sedrick Courtney case; mitochondrial DNA exclusion of convicted individual Lynette White case; familial DNA searching “Grim Sleeper”; familial DNA searching Shirley Duguay case; analysis of domestic animal hair continued



161



162



Appendix B: Selected DNA issues and case examples



Chapter



Case



2 4



5 1 2 3



5 7



3 4 3



Statistics Twins Ayturgrul v. The Queen; mitochondrial DNA Patrick Waring case; inclusion and exclusion Timothy Durham case; mixture erroneously treated as single source Josiah Sutton case; mixture interpretation Case from author’s files; complex mixtures Scott Peterson case; mitochondrial DNA statistics



2 5



7 1



Partial profiles and low levels Murder of Meredith Kercher Omagh bombing; low level DNA



2



6



Time of deposition Case from author’s files



7 8 1 2



Relevant evidence Murder of Meredith Kercher Frank Button case Damilola Taylor case Case from author’s files



2 3



Description



1 2 5



Poor collection practices and compromised evidence 2 O.J. Simpson case 7 Murder of Meredith Kercher 1 Omagh bombing



2 8



4 1 2 3 4



8



5 6



Contamination Farah Jama case; contamination during medical examination Adam Scott case; contamination during extraction step of analysis Jaidyn Leskie case; contamination between samples from different cases Murders of Margaret and Seana Tapp; contamination between samples from different cases Phantom of Heilbronn; contamination of sample consumables Errors Dwayne Jackson case; switching of DNA samples in vials Gareth Williams case; typographical error



Appendix C: Steps in review of evidence



C.1 Obtaining further information from documents and client One reason for requiring further information may be the possibility of innocent contact or other transfer that may explain a DNA match.



C.2 Review of scientific reports In the course of a review, certain questions should be asked in an effort to clarify the report contents: What is the nature of a sample? Is it blood, semen, cellular material? Has the nature of the material or the body origin been confirmed through confirmatory testing? Is the sample relevant to the matter in contention? What is the quality level of the sample? How many pertinent DNA results are there? Is there a possibility or probability of contamination? Define the strengths and limitations of the evidence even if they are not detailed in reports. 163



164



Appendix C: Steps in review of evidence



C.3 Requesting materials for review The case file notes (see Figure 3.6 in Chapter 3) should include the following items but a specific case may require additional documentation. Formulation of hypotheses Testing plan and sampling rationale Photographs or sketches taken in the laboratory Examination or “bench” notes made by laboratory analyst Electropherograms of DNA profiles including reference profiles Designations of alleles in electropherograms Thresholds used in designation of alleles Statistical calculations and description of population used Conclusions reached and bases for conclusions Standard forms related to testing such as batch numbers of reagents used, method, and equipment Communications between the analysts and others involved in the case Checks of results against staff DNA database Quality assurance (technical and administrative) reviews All case reports issued by laboratory including preliminary reports or notes of verbal examinations Certain documents that may not be included in a case file but may be needed for review and possible future use: Laboratory protocols such as DNA interpretation standards and statistical methods Records relating to the chain of custody Details about storage of materials Descriptions of software programs Equipment maintenance and calibration records Analyst training and proficiency test records Unexpected results or corrective action reports Quality assurance reports and audits



C.4 Examination by independent expert Always consult with your scientific expert before requesting material from the opposing party. The expert should examine all reports and raw data. He or she should also clear up any potential bias on the part of the opposing party’s expert.



Index A ABO blood grouping, 5, 16, 33 Acquittals, 18, 27, 29, 39, 134 Admissibility, 7, 8, 9, 69, 123 Adversarial systems, 144 Aitkin, C., 65 Alleles, 1, 17; see also Electropherograms appearance, 66 drop-in/out, 83, 85, 88, 99 frequencies, 68, 153 nonconcordant, 99 off-ladder, 57 reading tables of, 47, 48, 49, 50 Amelogenin, 17, 47, 155 AmpFLSTR Identifiler™, 22, 129 AmpFLSTR Profiler Plus™, 22 AmpFLSTR SGM Plus™, 22 Amplification, 53, 84, 90, 129 controls, 51 procedure, 53 sensitivity threshold, 84 with small amounts of DNA, 81, 84, 99 and stochastic variation, 86 Analytical threshold, 60–61, 96, 99 Ancient DNA studies, 14, 24, 117, 125 Andrews case, 7 Animal hair, 128–129 Archaeological findings, 118, 125 Armed robbery case, 26–27 Artifacts, 55, 57, 58, 59 dealing with, 88 pull-up, 57 stutter, 55, 56 Audit trails, 133, 135, 148 Australia aboriginal population, 26, 77 database, 22



Fraser review, 86, 88 legal cases in, 9, 20–21, 26, 29, 41, 73–74, 123, 124–125 Mungo man study, 125 Autosomal STR profiling, 4, 17, 45–46, 107, 114; see also DNA analyses differential extraction, 94, 107 inconclusive results with, 51, 74, 76, 105–106 main steps in, 45 statistical significance with, 70 Autosomes, 114, 129 Avery, Oswald, 2 Ayturgrul v. The Queen, 2012, 69, 123



B Bacteria, 129 Balding, D.J., 69, 76 Base pairs, 3, 46 Bases, 5 Bayes theorem, 68–69, 112 Beamish case, 128 Belfast bombing, 84 Bias, 74, 76, 98, 142, 143 Binary matches, 71 Biological model, 91 Biologist Special Advisory Group (BSAG) -Australia, 137 “Birthday problem,” 24, 71 Bite marks, 14–15 Blind testing, 139 Blood evidence, 34, 37–38, 41 confirming, 36 cross-examination questions for, 152 degradation, 38, 118–119 locating, 36 reference samples, 44



165



166 Blood grouping, 5, 16, 33 Blood relatives, 68, 77–78, 109 Bloodsworth case, 10 Body fluid testing, 36 Bone testing, 7, 117–119, 124, 125 Brother defense, 77 Buccal swabs, 44, 52 Buckleton, John, 70, 86, 88, 144 Burnt remains, 44, 119 Butler case, 109 Button case, 29, 133–134, 150



C Capillary electrophoresis, 51, 54 Case files, 47, 61, 62, 153 Castro case, 7 Cats, 128–129 Chain of custody, 149, 153 Chemoluminescence, 36 Chromosomes, 2, 5, 17 Clayton, T., 70, 98 Clothing samples, 34, 41, 134–135 Co-ancestry coefficient, 77 Cold cases, 134, 137 re-examining exhibits, 19, 126, 131, 150 solved through DNA, 18, 40, 41–42 Collection contamination during, 93, 133 instruments and transfer, 136, 137 personal protection equipment, 149 procedures, 8, 29, 84, 93, 149, 150 Combined DNA Index System (CODIS), 15, 22, 46, 101 Combined probability of exclusion (CPE), 69–70; see also Exclusion Combined probability of inclusion (CPI), 69 Computer software, 60, 70, 95, 153 Confessions, 6, 15, 107–108 Consensus strategies, 86, 91 Contamination; see also Transfer burden of proving, 135 between cases, 132, 134 cleaning techniques, 59, 124, 125 of clothing exhibits, 134–135 in cold cases, 131 documenting, 153 due to collection practices, 93, 133 identifying source of, 135 in laboratory, 133, 134 legal cases involving, 20–21, 93, 132–133, 134, 135, 136–137



Index with low level DNA, 92–93 of mitochondrial DNA samples, 124–125 prevention, 135 in rape cases, 131, 134, 135 sources of, 135, 136 Courtney case, 19, 121 Crick, Francis, 5 Crime scenes; see also Collection; Contamination mistakenly linking, 136–137 outdoors, 81 security, 149 transfer by personnel, 133, 136 CrimTrac National Criminal Investigation Database (NCIDD), 22 Cross-examination for difficult profiles, 152 of expert witnesses, 153 general questions, 151–152 for single source DNA of specific body fluid, 152 “CSI effect,” 30



D Dandruff, 40 Data files, 153 Data sharing, 18 Database searches, 23, 132, 133 Databases, 15, 71, 110; see also Populations comparison types, 23 core STR loci, 22 information stored in, 22 international, 123 national, 22 size, 68, 76 validation, 70 Death row, 10, 15, 122 Debriefings, 144, 151 Declared contributors, 75–76 Deen case, 72 Defense experts for, 148–149, 151 fallacy of, 71–73 hypothesis, 67 questioning DNA evidence, 30, 148–149 responsibilities, 145 Degradation, 19, 24 of blood evidence, 38, 118–119 exposure to elements, 18, 81, 119 factors affecting rate of, 18



Index fire damage, 119 observed on electropherograms, 57, 58, 82 resulting in partial profiles, 81 of saliva, 40 “ski slope” effect, 81–82 Deoxyribonucleic acid (DNA) molecule, 33 in cellular makeup, 16 and concept of individuality, 4–5 and law of inheritance, 1 structure, 2–3 variable numbers of tandem repeats, 5 Deposition; see also Transfer modes of, 26, 137, 148, 150 through legitimate contact, 132, 150 time of, 25–27, 43, 149 Dewey case, 18, 19 Dinosaur DNA, 125 Disclosure, 147, 153 Discovery requests, 153 DNA, see Deoxyribonucleic acid (DNA) molecule DNA analyses; see also Amplification; Interpretation automation of, 140 batch processing, 51, 60, 132 of bone, 117–119 case documentation, 61, 62 confirmatory tests, 34, 36, 138 controls, 51–52, 149 detection, 54 extraction, 52, 94 identifying body fluid, 34, 36 objectives of, 35–36 peak designation, 60–61 processing steps, 51, 132 quantification, 52–53 quantity of extract, 151 retesting samples, 143–144 sensitivity, 90 separation, 54 from single cell, 10, 25 technical issues, 55, 56, 57, 58, 59 template thresholds, 84, 86, 90 time required for, 59–60 DNA evidence; see also Collection; Contamination; Degradation; Deposition; Transfer admissibility, 7, 8, 9, 69, 123 bagging, 138, 149 confirmatory tests, 34, 36 in context, 24–25, 131



167 decontamination, 124–125 endogenous, 125 mishandling, 8, 27–28, 84, 133, 140 persistence of, 25–27 preservation, 18 quantity, 10, 25, 34–35, 37, 82, 84 querying at trial, 148–149 relevance, 9, 25, 27–28, 29, 37, 38–39 retention/storage, 18, 143–144 searching for, 35–37 as sole evidence in case, 131, 132, 133, 140, 149 stability of, 18–19 testing all collected evidence, 29, 37, 133–134, 150 of unspecified body origin, 34, 93, 138, 148, 152 value of, 9 DNA extract, 18, 143, 151 DNA fingerprinting, 5, 7 DNA mixtures analysis steps, 98 complex, 95, 100–101 contributors to Y-STR profile, 113 cross-examination questions, 152 discussion between opposing counsel, 148 four or more contributors, 95, 100–101 interpretation, 96, 98, 113, 138, 142–143 legal cases involving, 75, 94, 95–96, 100 likelihood ratios, 97, 99, 100, 113–114 low template, 99 major and minor contributors, 28, 95, 97, 113, 124 male and female, 106, 107 of mitochondrial DNA, 124, 125–126 mixture ratios, 113–114 two or more people, 93–97 unknown number of contributors, 98 unrestricted versus restricted calculations, 97 victim DNA, 36 DNA profiling, 5, 7, 9; see also DNA analyses; Mitochondrial DNA (mtDNA); Y-STR profiling advantages of, 14, 76, 148 biological materials allowing, 33–35 discrimination power, 6, 14, 15–16, 65 fallibility, 13, 30 genetic basis for, 16–18 kits, 17, 22, 23, 84, 129 misconceptions, 3, 144



168



Index



public perception of, 3, 15 techniques used, 4, 17, 105, 128, 129, 149 DNA replication, 5 DNA sequencing, 3–4 DNA technology early court challenges to, 7–9 improvements in, 9–10 overreliance on, 140–141 Documentation, see Case files; Reports Dogs, 125, 128, 129 Dominant genes, 1 “Dragnets,” 23 Drop-in/out, 83, 85, 86, 88, 99 Duguay case, 128 Duke of Edinburgh, 117 Durham case, 75



Exhibits; see also DNA evidence relevance, 29 searching for DNA on, 35–37 transfer between, 138, 149 Exonerations, 10, 13, 15, 73, 96, 139 Experts, 30, 147 alternative theories, 73, 98, 147 conflicting conclusions by, 143 cross examination questions, 153 explaining limitations, 73, 131, 144 obligations of, 131, 144 subjectivity, 141–143 when to call to testify, 148 Eye color, 44



E



“Fallacy of the transposed conditional,” 71–73 False positives, 75, 138, 139 Familial matching basis for, 127 legal cases, 126–127, 127–128 with mitochondrial DNA, 126–128 permission policies, 127 of Romanov family, 117 search technique, 24, 126, 127 Feces, 43 Fiber evidence, 144 Fingernail scrapings, 40 Firearms analysis, 127, 136 Forensics, 13–14, 30 Franklin, Rosalind, 2, 3 Franklin case, 127–128 Fraser, Jim, 86, 88, 144 Fraser review, 86, 88 Frequencies, see Statistics Full disclosure, 147, 151



Electropherograms artifacts, 55, 56, 57, 58, 59 experiment involving, 88–90 indications of degradation, 57, 58, 82 measurement units, 54 partial profiles, 58, 83 peak designation, 60–61, 66, 89, 99 reading, 54–55 samples, 49, 50, 83, 87, 89 “ski slope” effects, 59, 81–82 stochastic threshold, 47, 61, 90 Electrophoresis, 51, 54 Emerging techniques, 43–44 EMPOP 9 database, 123 Enhanced interrogation approach, 86 Enhancement techniques, 84, 90 Enzyme typing, 5 Epithelial cells, 34, 36, 41 Error rates, 138–140 Error types, 139 Ethnicity, 68, 77 European Network of Forensic Science Institutes (ENFSI), 137 Evett, Ian, 76 Evidence, see DNA evidence Exclusion, 9, 14, 15–16, 22, 149 combined probability of, 69–70 expert disagreement on, 143 percentage, 123 standards for, 142 as used in forensic statements, 73–75 with Y-STR profiling, 110



F



G GeneMapper®, 60 Genes, 1, 2, 3, 4, 17 Genescan®, 54 Genomes full human, 3–4 mitochondrial versus nuclear, 119 Genotype frequencies, 67, 68, 70 Genotyper®, 54, 60 Genotypes, 17, 46, 67 German Stain Commission, 101



Index Gesah case, 134–135 Gill, Peter, 6, 86, 88, 144 Gloves, 149 Goldman v. Simpson, 8 Gosling, Ray, 2 “Grim Sleeper” case, 127–128



H Hair color prediction, 44 degradation, 19 from domestic animals, 128–129 evidence, 40 mitochondrial DNA profiling, 120, 121, 122 reference samples, 44 Haplotypes, 105, 108, 109 cited in report, 149 frequency estimates, 70, 110–111, 111–113 meaning of Y-STR match, 111–113 surveying method, 112 Hardy–Weinberg equilibrium, 68 Hearsay, 9 Heilbronn phantom case, 93, 136–137 Heredity, 1, 17, 77 Heteroplasmy, 120, 121 Heterozygote, 61, 97 Heterozygous, 1, 17 HIrisplex test, 44 Homozygous, 1, 17 Human genome project, 3–4 Human hairs, see Hair Human remains identification, 24, 117, 118, 119 Hypotheses, 13, 36, 67, 133



I Identical twins case, 21 Identity, 14, 17, 71 Inbreeding coefficient, 68, 78 Inclusion combined probability, 69 standards for, 142 as used in forensic statements, 73–75 with Y-STR profiling, 110 Inconclusive findings, 51, 74, 76, 143 Inheritance, 1–2, 17 Inhibition of DNA, 59, 81, 82, 83 Innocence Project, 7, 10, 13, 75, 109, 121



169 Insects, 43 Intelligence-led policing, 23–24 International Society of Forensic Genetics (ISFG), 70, 92 Interpretation, 66, 71 analytical threshold, 60–61, 96 expert disagreement, 143 with low level DNA, 85, 86, 87 with mixtures, 96, 98, 113, 138, 142–143 objectivity, 14, 141–143 prior to reference sample comparison, 97, 98, 142 subjectivity, 66, 141–143 uninterpretable profiles, 101



J Jackson case, 140–141 Jama case, 20–21, 135, 141–142, 144, 149 Japan case, 107–108 Jefferson, Thomas, 108 Jeffreys, Alec, 5, 6, 45 Junk DNA, 4 Juries considerations of, 72, 73 and CSI effect, 30 decisions by, 8, 9, 27, 42, 100 information conveyed to, 30, 84, 96, 145



K Kercher murder case, 27–28, 93, 135, 138 Knox case, 27–28, 91 Krone case, 14–15



L Laboratories; see also Reports accreditation, 55, 61, 153 audits, 132 background DNA levels, 135–136 contamination in, 133, 134, 136 error, 8 process reviews, 61 protocols, 153 quality control, 132–133, 140 records, 151 sterility versus DNA-free materials, 136, 137 test validation, 55, 132 Ladders, 54



170 Law of Independent Assortment, 2 Law of Segregation, 1 Legal cases absence of evidence, 15 collection practices, 8, 27–28, 84 compromised evidence, 8, 27–28, 84 confessions, 6, 15, 107–108 contamination, 20–21, 132–133, 134, 135, 136–137 DNA from maggots, 43, 44 DNA stability, 18–19 domestic animal hair, 128 errors, 140–141 familial DNA searching, 126–127, 127–128 first murder solved by DNA, 6 involving mixtures, 75, 95–96, 100 low levels, 27–28, 84 mass screening, 6, 108–109 mitochondrial DNA, 120, 121 partial profiles, 27–28, 84 precedent setting, 128 questionable science, 14–15 relevant evidence, 27–28, 29, 37, 38–39 statistics, 21 time of deposition, 26–27 with touch DNA, 42–43 use of database to solve cold case, 42 warning signs, 131, 149–150 with wearer DNA, 41–42 Legal fallacies, 71–73 Leskie case, 134 Likelihood ratios, 20 approach, 66–68 basis for, 67 with mixtures, 68, 97, 99, 100, 113–114 versus probability of exclusion, 69–70 sampling uncertainty, 68, 76 with stochastic effects, 88 verbal descriptor scale, 76 Linkage equilibrium, 68 Locard’s Theorem, 25 Loci, 17 LoComatioN, 97 Locus drop-out, 83, 85 Low copy number (LCN), 84, 107; see also Low level DNA Low level DNA contamination, 92–93 cross-examination questions, 152 definitions, 82, 84–85



Index detection protocols, 85 enhancement techniques, 84, 90 experiment using, 88–90 interpretation of profiles, 85, 86, 88 legal cases, 27–28, 84, 138 mixtures, 99 and partial profiles, 81, 82 reliability of results, 82, 90–92 statistical models, 91–92 stochastic effects, 85–86, 87, 88 Low template DNA, see Low level DNA



M Maggots, 15, 43, 44 Masking, 66, 85, 88, 92, 97 Mass disasters, 24 Mass graves, 117, 119 Mass screenings, 6, 108–109 Matches, 20; see also Likelihood ratios; Probabilities binary, 71 coincidental probability, 111–112 in databases, 18, 22, 23, 71, 127 no-match systems, 71 partial, 126 reporting, 111–113 subjectivity in, 142 with Y-STR profiling, 109, 110 Maternal inheritance, 117, 120 Mendel, Gregor, 1 Mendelian inheritance, 1–2, 68, 77, 126 Mengele, Josef, 7 MeowPlex kit, 129 Messenger RNA (mRNA), 36 Microbial forensics, 129 Microsatellites, 6, 17 Minifiler STR kit, 129 Minisatellites, 5, 6 Missing person cases, 18, 45 Mitochondria, 119–120 Mitochondrial DNA (mtDNA) profiling, 4, 17, 24, 40, 149 analysis of bone, 117–119 contamination of samples, 124–125 counting method, 111, 112, 123 disadvantages of, 120 discrimination of results, 120, 123 familial searching, 126–128 from hair, 120, 121 and heteroplasmy, 120, 121



Index inheritance, 117, 120 international database of, 123 legal cases involving, 69, 120, 121, 122, 124–125, 126–128 mixture profiles, 124, 125–126 regions, 121 relevant population, 122 statistical analysis, 122–123 survival, 120 versus Y-STR profiling, 120 Mixtures, see DNA mixtures Multiplexes, 45, 46, 120 Mutations, 109, 110, 121



N Nasal secretions, 34, 40 National Criminal Investigation Database (NCIDD), 22 National Institutes of Health, 4 Natural disaster victims, 24 Netherlands Forensic Institute, 139, 140 New York Medical Examiner’s Office, 109 New York Supreme Court, 7 Nichols, R.A., 69, 76 No-match systems, 71 Nobel Prize, 2, 3 Novel techniques, 129 Nuclear DNA, 40 in cells, 17 profiling using, 17, 149 of Romanov family, 117 Nucleotides, 4



O Objectivity, 14, 141–143 O.J. Simpson trial, 8 Omagh bombing, 84, 93



P Paint evidence, 144 Partial matches, 126 Partial profiles; see also Low level DNA causative factors, 81 cross-examination questions, 152 electropherograms of, 58, 83 legal cases, 27–28, 84, 141 statistical value, 81 Paternal inheritance, 109, 110, 117, 118



171 Paternity testing, 108 Patrilocality, 110 Pauling, Linus, 2 PCR, see Polymerase chain reaction (PCR) Peer reviews, 61, 132 People v. Castro, 7 People v. Simpson (1995), 8 Personal belongings, 45 Personal protective equipment, 149 Peterson case, 122 Phantom of Heilbronn, 93, 136–137 Phenotyping, 44 Pitchfork case, 6 Plant material, 129 Plea bargains, 9, 143 Poland case, 108–109 Polymerase chain reaction (PCR), 6, 18, 45 number of cycles, 53, 82 purification post-PCR, 90 reduced volume, 90 Populations and ethnicity, 68, 77 relevance, 77, 110–111, 122 sample size, 68, 69, 110 PowerPlex 16™, 22 Presumptive tests, 36, 59, 62 Pretrial hearings, 7, 8, 144 Pretrial reviews, 151 Privacy, 23, 127, 128 Probabilities approach, 19–21, 66–67, 91–92, 144 of coincidental matches, 111–112 conditional, 68 of exclusion, 69–70 “fallacy of the transposed conditional,” 71–73 of guilt/innocence, 71, 72 Product rule, 68, 77 Propositions, 67–68, 77, 98 Prosecution calling experts, 148 fallacy of, 71–73 hypothesis, 67 questioning DNA evidence, 148–149 responsibilities, 145 Puch-Solis, R., 65 Pull-up, 57



Q Quality control, 132–133, 140, 148



172



Index



R



S



R. v. Deen, 1994, 72 R. v. Doheny and Adams (1997), 8 Race, 67, 77 Random man not excluded (RMNE), 70 Random match probability (RMP), 66–67, 99 Rape cases, 7, 42–43, 107, 108, 111 Bloodsworth case, 10 Butler case, 109 Button case, 29, 133–134, 150 Dewey case, 18, 19 DNA mixtures, 74, 75, 93, 94, 143 Durham case, 75 evidence collection, 39, 133 evidence contamination, 131, 134, 135 fought on consent grounds, 25 gang rape, 74, 113, 143 Jama case, 20–21, 135, 141–142, 144, 149 penetration without DNA deposition, 39 Pitchfork case, 6 postcoital interval, 106, 107 Scott case, 132–133, 149 Thibodeaux case, 14–15 Ware case, 120 Waring case, 73–74 Rarity, 20, 71 Reagent blanks, 51, 52 Recessive genes, 1 Reference samples, 44–45, 135 and bias, 98, 143 comparison to crime sample, 66, 98 Relative fluorescent units (RFUs), 54, 99 Relatives, 21, 68, 77–78 Reliability, 82, 90–92 Repeat offenders, 23 Reports; see also Case files declared contributors, 75–76 differing hypotheses, 133 inclusion and exclusion, 73–75 terminology, 76, 84, 112, 113 testing rationale, 133 Restriction fragment length polymorphism (RFLP), 5, 6, 45 Retesting, 143–144 RNA, 36 Romanov family, 117, 118, 121 Royal Statistical Society, 65 Russian royal family, 117, 118



Safeguards, 133 Saliva, 34, 39–40 cross-examination questions for, 152 DNA degradation, 40 electropherogram of, 89 Sampling correction, 68, 69, 76 Sampling errors, 112 Science discipline of, 13 exactness, 144 philosophy of, 144 Scientific acceptance, 7 Scientific method, 13, 36, 132, 133 Scientific Working Group on DNA Analysis Methods (SWGDAM), 78, 97, 98, 112, 137 Scientists, see Experts Scott case, 132–133, 149 Screening tests, 36, 59, 94, 106 Semen evidence, 29, 34, 38–39; see also Rape cases cross-examination questions for, 152 differential extraction, 38, 107 location in context, 38–39 masked stains, 36 sperm survival, 38 testing all available evidence, 150 Serology tests, 5, 16, 33 Sex chromosomes, 2, 129 Sex markers, 17 SGM Plus™ system, 47 “Shedders,” 43 Short tandem repeats (STR), 17–18, 20, 45–46; see also Autosomal STR profiling; Y-STR profiling frequency tables, 153 markers, 17 primers, 53 Siblings, 77–78 Simpson case, 8, 70, 133 Size bias, 76 Skin cells, 34, 40, 43 Smith case, 42 “Snaggle tooth killer,” 14–15 Software, 54, 60, 70, 95, 153 Spermatozoa, see Semen evidence State v. Woodall (1987), 7



Index Statistics, 19–21; see also Likelihood ratios; Probabilities agreement on use of, 147 assumptions, 71, 112, 144 Bayesian approach, 68–69, 112 biological model, 91 combining autosomal and Y-STR, 114 confidence intervals, 112 conservativeness, 92, 97, 112 counting method, 111, 112, 123 exclusion percentage, 123 frequencies, 68–69, 110–111, 112, 113, 123 guides to, 65 in mitochondrial DNA analysis, 122–123 with mixtures, 97 models, 91–92, 97 “2p rule,” 92 product rule, 68, 77 random data, 142 reporting method, 123 sampling errors, 112 short cut calculations, 92 significance, 66, 69, 70, 85, 121 understanding, 65 weights, 110 of Y-STR profiling, 110–113 Stochastic effects, 82, 85–86, 88 accounting for, 90 in complex mixture, 101 electropherograms with, 87 Stochastic threshold, 47, 61, 90 STR, see Short tandem repeats (STR) Stutter, 88 Subjectivity, 66, 141–143 Suboptimal profiles, 66, 82; see also Low level DNA Suspect screening, 23 Sutton case, 95–96, 139 Sweat, 41



T Tapp murders case, 134–135 Targeted operations, 23 Taylor case, 37 Technology, see DNA technology Teeth, 118, 119, 125 Terminology, 76, 84, 112, 113 Testimony, see Experts Texas Sharpshooter Fallacy, 142



173 Thibodeaux case, 14–15 Thompson experiment, 88–90 Touch DNA, 34, 138, 148, 151 case involving, 42–43 mixtures, 93 Trace evidence, 25, 34; see also Touch DNA Traits, 1, 2 Transcription errors, 140, 141 Transfer case involving, 26–27 between cases, 132, 134 by crime scene personnel, 133 between exhibits in same case, 138, 149 to firearms, 136 modes of, 73 to packaging, 138 during police searches, 136 primary, 26 secondary, 25, 26 solving cold cases, 19 two-way, 27, 131, 150 unintended, 153 Trial preparation, 144 TrueAllele, 97 Twins case, 21



U Unbalanced peak heights, 88 Uncertainty, 76 United Kingdom (U.K.) database, 22 legal cases in, 6, 126–127 as pioneers of familial DNA searching, 126 review of interpretation with low level DNA, 85 Unrestricted combinatorial method, 97 Urine, 43 U.S. 13-locus CODIS system, 101 U.S. Department of Energy, 4



V Vaginal secretions, 40–41 Vaginal swabs, 39, 93 Valentine’s Day murder, 126–127 Validation studies, 20, 61, 86, 132 Verbal descriptors, 76 Victim DNA, 15, 36, 47, 93, 150



174 Vincent, Justice F., 20, 21, 135, 144 Viruses, 129 Voir dire examination, 9



W Ware case, 120 Waring case, 73–74 Watson, James, 2, 3, 5 Wearer DNA, 34, 41–42 Weir, Bruce, 76 Wellcome Trust, 4 West Virginia Supreme Court, 7 White case, 126–127 Wilkins, Maurice, 2 Williams case, 141 Woodall case, 7 Wrongful convictions, see Innocence Project



Index Y Y-chromosome, 109 Y Chromosome Haplotype Reference Database (YHRD), 110 Y-STR profiling, 4, 17, 105–106 benefits of, 106–107, 108, 109 cases with, 107–108, 108–109 coincidental match probability, 111–112 combining with autosomal STR, 114 frequency estimates, 110–111 mass screening through, 108–109 match report, 111–113 mixture ratios, 113–114 versus mtDNA profiling, 120 number of male contributors, 107, 113 statistics of, 110–113 theory of, 109 weakness of, 109