Human Identification: The Use of DNA Markers PDF
Preview Human Identification: The Use of DNA Markers
HUMAN IDENTIFICATION: THE USE OF DNA MARKERS Contemporary Issues in Genetics and Evolution VOLUME4 The titles published in this series are listed at the end of this volume. Human Identification: The Use of DNA Markers Edited by BRUCE S. WEIR Contributions with an asterisk in the table of contents were first published in Genetica, Volume 96 no. 1-2 (1995) Springer-Science+Business Media, B.V. Library of Congress Cataloging-in-Publication Data Human 1 dent if 1 cat 1 on the use of DNA markers I edited by Bruce S. Weir. p. em. -- <Contemporary tssues in genetic and evolut1on ; v. 4) Includes btbltographtcal references and index. ISBN 978-94-017-1803-5 ISBN 978-0-306-46851-3 (eBook) DOI 10.1007/978-0-306-46851-3 1. DNA fingerprinting. I. ~etr. B.S. <Bruce S.). 1943- II. Series. RA1057.55.H85 1995 614' .1--dc20 95-17081 Printed on acid-free paper All Rights Reserved © 1995 Springer Science+Business Media Dordrecht Originally published by Kluwer Academic Publishers in 1995 Softcover reprint of the hardcover 1st edition 1995 No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the copyright owner. Contents * Introduction by B.S. Weir * D.J. Balding and R.A. Nichols. A method for quantifying differentiation between populations at multi-allelic loci and its implications for investigating identity and paternity 3 * J.F.Y. Brookfield, The effect of relatedness on likelihood ratios and the use of conservative estimates 13 * B. Budowle, The effects of inbreeding on DNA profile frequency estimates using PCR-based loci 21 * R. Chakraborty and Z. Li, Correlation of DNA fragment sizes within loci in the presence of non- detectable alleles 27 * A. G. Clark, J.F. Hamilton and O.K. Chambers, Inference of population subdivision from the VNTR distributions of New Zealanders 37 * R.N. Curnow, Conditioning on the number of bands in interpreting matches of multilocus DNA profiles 51 * P. Donnelly, Match probability calculations for multi-locus DNA profiles 55 * P. Gill and I. Evett, Population genetics of short tandem repeat (STR) loci 69 * D.W. Gjertson and J.W. Morris, Assessing probability of paternity and the product rule in DNA systems 89 * D.H. Kaye, The forensic debut of the NRC's DNA report: population structure, ceiling frequencies and the need for numbers 99 * K. Lange, Applications of the Dirichlet distribution to forensic match probabilities 107 * R. Lempert, The honest scientist's guide to DNA evidence 119 * P.J. Maiste and B.S. Weir, A comparison oftests for independence in the FBI RFLP data bases 125 * N.E .. Morton, Alternative approaches to population structure 139 * B. Robertson and G.A. Vignaux, DNA evidence: wrong answers or wrong questions? 145 * W.C. Thompson, Subjective interpretation, laboratory error and the value of forensic DNA evidence: three case studies 153 * D. Zaykin, L. Zhivotovsky and B.S. Weir, Exact tests for association between alleles at arbitrary numbers of loci 169 B.S. Weir, A bibliography for the use of DNA in human identification 179 * Contributions indicated with an asterisk were first published in Genetica, Volume 96 no. 1-2 (1995). B.S. Weir(ed.), Human identification: The Use of DNA Markers, 1-2, 1995. © 1995 Kluwer Academic Publishers. Introduction As this volume goes to press it is not known whether O.J. Simpson will be convicted of having murdered his former wife and her friend. It is clear, however, that DNA evidence is playing a central part in the trial. The publicity given to the trial has meant that this kind of evidence has been discussed widely in the media, and so may mark a turning point-in the future there will be an expectation that DNA evidence will be sought in the same way that crime scenes are routinely examined for fingerprints. In documents filed with the court, the Simpson defense team challenged the admissibility of DNA evidence on several grounds, including population genetics and statistics. These issues all relate to the meaning to be attached to the finding of a match between evidentiary samples and blood taken from a suspect or a victim. The more rare the matching DNA profile is in the relevant population, the more likely it is that there is a common source for the matching samples. Estimating the degree of rarity is the main theme of the papers in this volume. Each of the contributors to the volume has been involved in the debate over the use of DNA for human identification, and although a few of the major players declined to contribute, the volume represents the diversity of current views. Although contributors were selected to provide alternative approaches, there was also an attempt to select authors who have used sound scientific reasoning and have displayed some balance in their presentations. There is no doubt that challenges to statistical and population genetic methods applied to DNA profile frequencies have led to strengthening of those methods. It is to be hoped that this volume will show past challenges to have largely been met, and to indicate what remains to be done. There are going to be situations, such as those requiring discrimination between relatives or members of small populations, where statistical issues will continue to be important but it may be that biology eventually overtakes statistics. There may well come a point where so much of the genome is examined, to determine a DNA profile, that there can be little doubt of identity. As with all forms of evidence, of course, the possibility of error or fraud may need to be considered. A detailed synopsis of each chapter will not be given in this Introduction. I have chosen instead to add a few comments at the end of each paper. It should be mentioned that each paper went through the usual peer-review process, and that each paper was revised accordingly. All the contributors are grateful to the reviewers. Not only is the outcome of the Simpson trial unknown at this point, but so is that of the second National Research Council (NRC) committee deliberations. The debate over the use of DNA in human identification rose to such a vocal level in early I 990 that the NRC convened a panel of distinguished scientists to examine the issues. It may not have been fully appreciated that the debate was taking place almost entirely in the courtroom, where approximately equal numbers of experts for each side were engaged. The scientific literature, a more fitting forum for scientific discourse, was relatively thin. Since that time, there has been a large literature and publication of many studies of forensic DNA frequency databases. In the normal course of events this literature would have served to shed light on the debate. Instead, attention was shifted to the NRC report that was published in 1992. Although the recommendations of the report, especially the so-called ceiling principle, were motivated by the desire to avoid courtroom battles, they often had the opposite effect. Some courts were eager to embrace a set of guidelines issued under the aegis of the National Academy of Sciences, but courts did not realize that scientists are more likely to be guided by the scientific literature than expert reports. It is expected that the second report will both avoid statistical errors and refer to published studies of data. It would be a pity if this volume were seen merely as a response to a debate over how to calcu late the frequency of a DNA profile. Instead, it should be regarded as a celebration of the power of molecular biology to identify the possible contributors to biological samples. The benefits of this technol ogy are considerable. Determining paternity in disputed cases is probably the major use of DNA profil ing. There are considerable benefits to society of ensuring that children are adequately supported. Few er people are affected, but society also gains when violent criminals are identified and convicted-as 2 well as when those who are wrongly suspected or wrongly convicted are exonerated by DNA profiling. Smaller numbers still are involved when identification is needed following major disasters such as war or airplane crashes, but identification of remains brings comfort and closure to friends and family. Human identification by means of DNA is playing a large role in evolutionary studies, aided by the ability to extract DNA from ancient remains. Profiling non-human species also serves human welfare. Protection of proprietary crop species or of endangered animals is aided by DNA identification, as is quality control of cell cultures. The new analyses presented in this volume will go a long way to ensuring that DNA identification is made on a sound scientific basis, so that the many benefits of such identification will follow. Raleigh, N.C. B.S. Weir B.S. Weir (ed.), Human Identification: The Use of DNA Markers, 3-12, 1995. 3 © 1995 Kluwer Academic Publishers. A method for quantifying differentiation between populations at multi-allelic loci and its implications for investigating identity and paternity David J. Balding & Richard A. Nichols School of Mathematical Sciences and School of Biological Sciences, Queen Mary & Westfield College, University ofL ondon, Mile End Road, London El 4NS, UK Received 16May 1994 Accepted26July 1994 Key words: DNA profiles, paternity, Wright's FsT. coancestry, forensic science Abstract A method is proposed for allowing for the effects of population differentiation, and other factors, in forensic inference based on DNA profiles. Much current forensic practice ignores, for example, the effects of coancestry and inappropriate databases and is consequently systematically biased against defendants. Problems with the 'product rule' for forensic identification have been highlighted by several authors, but important aspects of the problems are not widely appreciated. This arises in part because the match probability has often been confused with the relative frequency of the profile. Further, the analogous problems in paternity cases have received little attention. The proposed method is derived under general assumptions about the underlying population genetic processes. Probabilities relevant to forensic inference are expressed in terms of a single parameter whose values can be chosen to reflect the specific circumstances. The method is currently used in some UK courts and has important advantages over the 'Ceiling Principle' method, which has been criticized on a number of grounds. 1. Introduction es, such as those involving partial profiles or large numbers of possible culprits, many of whom share the The genetic composition of human populations varies defendant's ethnic background. However, the foren because of, among other factors, their differing evo sic use of DNA profiles need not be invalidated as a lutionary histories and patterns of dispersal and inter consequence. One approach to allowing for popula breeding. The magnitude of the effect of this genetic tion differentiation, the 'Ceiling Principle', has been differentiation on the forensic evaluation of DNA pro proposed by the US National Research Council (NRC) file evidence is controversial. It is the practice of many (DNA Technology in Forensic Science, Nat!. Acad. forensic scientists to ignore coancestry except, pos Press, Washington D.C., 1992). The principle has been sibly, in cases where genetically isolated populations widely criticized (Robertson & Vignaux, 1992; Devlin, or close relatives are clearly involved. Some authors Risch & Roeder, 1993; Morton, 1993a; Weir, 1993a). argue, however, that uncertainty about possible lev In particular, the principle is inflexible and cannot be els of differentiation may invalidate such an approach adjusted to the circumstances of a particular case, in (Lewontin & Hartl, 1991; Krane et al., 1992). Others part because it incorporates the view that the defen take the view that typical levels of differentiation are dant's ethnicity is irrelevant to inference. We propose sufficiently small that they may routinely be neglected a method for quantifying the effect of genetic differ (Chakraborty & Kidd, 1991; Roeder, 1994). entiation in terms of a single parameter, which can We argue for an intermediate position: even small often be interpreted in terms of coancestry. Debates levels of genetic differentiation can be important and about the effect of population heterogeneity in partic the effect should not be ignored. To do so would unfair ular cases can thus be simplified to a discussion of val ly overstate the strength of the evidence against the ues for the parameter appropriate to the circumstances. defendant and the error could be crucial in some cas- Our proposed method has previously been described 4 (Balding & Nichols, 1994) and is currently used in profile in some population. This use of profile frequen some UK courts. Here, we develop the justification cies in place of the match probability is inappropriate for the method and extend its application to paternity for several reasons. The concept of 'match' clearly testing. The use of DNA profile evidence when incest involves two profiles, not one, and there seems no log is alleged in paternity cases is becoming increasing ical framework for linking profile frequencies with the ly common and the proposed method is particularly issue of the defendant's guilt or innocence, which is appropriate in such cases. the crucial issue in court. In particular, it is unclear how to allow coherently for the possibility that the culprit is related to the defendant, or shares ancestry 2. Key issues in forensic inference through common origin in a subpopulation. Perhaps most importantly, there seems no logical framework Although the literature on forensic identification using for combining the DNA evidence, quantified by a pro DNA profile evidence is now extensive, many funda file frequency, with the non-DNA evidence. mental statistical issues are still not widely appreciat Correct definition of the match probability clarifies ed. Balding and Donnelly (1995) consider the forensic much of the current debate. A general discussion of identification inference problem in a general setting 'reference populations' can be avoided and neither is it and their analysis clarifies several issues. In particu necessary to consider hypothetical 'random' selections lar, they show that the weight of evidence against the of suspects. Crucially, a coherent framework becomes defendant depends on, for each possible perpetrator available for incorporating the effects of shared ances other than the defendant, the ratio of the likelihood of try, on both recent and evolutionary timescales. Since the DNA profile data if he were the culprit, to its likeli match probabilities are conditional probabilities, they hood if the defendant were the culprit. These likelihood cannot be estimated directly from database relative fre ratios should then be summed by the jury, weighted by quencies. Correlations in profile possession must be their probability, based on the non-DNA evidence, that explicitly modelled in terms of population genetic the each possible culprit is the true culprit. ory in addition to the available data. Consequently, To facilitate the discussion, it is common to make the ethnicities of both defendant and possible culprits four simplifying assumptions: are relevant to inference. Some authorities ignore cor 1. that the crime sample DNA is that of the culprit; relations in profile possession and, instead, use 'con 2. that matches are unequivocal; servative' estimates of relative frequencies. The Ceil 3. that if the defendant were the culpritthen the defen ing Principle, for example, is based on this approach. dant and crime sample DNA profiles would be cer However unless the correlations are specifically taken tain to match; and into account, it is impossible to assess what level of 4. the fact that the defendant's DNA profile was inves 'conservativeness' is appropriate. tigated is not, in itself, informative about his/her Some of the current debate concerning population profile. differentiation focusses on statistical tests of hypothe These assumptions are not valid in general, but they ses of independence in forensic databases (Geisser & allow us to focus on other important issues and devia Johnson, 1993; Weir, 1993b). The tests are complicat tions from them can be addressed within the framework ed by the experimental difficulties involving apparent discussed here. See Balding and Donnelly (1995) for homozygotes. It is, in any case, difficult in princi further discussion. ple to draw conclusions relevant to forensic inference Under these four assumptions, each likelihood ratio from the outcomes of such tests. Population differ is simply the conditional probability that the possible entiation indubitably exists, the question of interest culprit has the profile given that the defendant has it, concerns the magnitude of its effect on match proba that is, the 'match probability'. Note that the match bilities. Failure to reject a null hypothesis of no differ probability may also be formulated in terms of the entiation reflects some combination of insufficient, or probability that the defendant has the profile condition inappropriate, data, low power against the alternatives al on the event that the alternative culprit has it, but we of interest and small magnitude of effect. Such tests find the former definition to be more convenient. are thus not directly helpful in forensic inference. We Many authors ignore the conditioning on the propose parameter estimation, both point and interval, observed profile and take the match probability to be as an alternative to hypothesis testing. equivalent to the relative frequency ofthe defendant's
The list of books you might like
