In 2009 D was convicted for the 1990 murder of C in C's own apartment. In a post‐conviction review DNA attributed only to D and C (and no others) was found underneath the fingernails of C. At trial, in a hearing for a retrial, and in a pending complaint to the Massachusetts Forensic Science oversight board, the statements of government witness' regarding the meaning of the DNA evidence at activity level were a topic of debate. In this paper, a Bayesian network (BN) evaluation of this evidence is presented. This BN uses the propositions that D was the attacker ( H p ) versus an alternate proposition that he was not the attacker ( H a ). The alternate, which was inferred from defense questioning, requires that transfer occurred from a social meeting 2 to 4 weeks earlier. The evaluation presented here suggests an LR of the order of 800. This analysis suggests that, while the original testimony may not have been prepared for in a formal manner, it was not misleading to a lay jury.

We have generated a general lookup table of LRs that could be used in the general case of a person of interest (POI) accused of attacking a complainant and the evidence is DNA from fingernails of the complainant. We consider that social interaction may have occurred between the complainant and the POI or others, comments from the complainant about scratching the attacker (or not), and visible scratches on the POI (or not). The LRs are generally small, ranging from a few thousand to less than one. If POI and no other DNA (except the complainant (C)) are present with no opportunity for social interaction between C and POI, then the LR is assigned a value in the thousands. The evidence supports the Hp proposition (LR between 3 and 10) more than the alternative if there has been socialization between C and POI, and as long as there is no unknown DNA observed. The presence of unknown DNA as well as POI DNA gives LRs from 500 to 700 in the absence of social contact between C and POI. The evidence is neutral if neither POI nor unknown DNA is present. DNA analysis of fingernails is exculpatory if the POI is excluded from the findings and unknown DNA is present. Considering whether the complainant claims to have scratched her attacker increases the discrimination power of the model by a small amount, increasing the support for Hp when only POI DNA is present, or increasing the support for Ha if only unknown DNA is present.

This report discusses the analysis of the evidence given activity level propositions for a murder case in Austin, Texas. The testimony in this case has been the subject of a complaint to the Texas Forensic Science Commission and a subsequent report. A Bayes' net is constructed for the propositions Hp: A took W's bike from the scene of the murder to an alley where her Jeep Grand Cherokee was parked and then threw the bike into some bushes. Ha: Somebody other than A took W's bike from the apartment, rode it, and disposed of it in some bushes. A has never had any contact with W's bike. Based on published data and explicitly stated subjective probability assignments, the likelihood ratio is about 1300 in favor of Hp. This analysis supports that the original testimony given in rebuttal was correct.

Evaluation of DNA given activity propositions (EGALP) is a widely discussed topic at this time, particularly in the United States. There is concern about opinions given in testimony that are not properly founded. Guidance on evaluative reporting given alleged activities can be found in published papers, official documents, and specialized textbooks. In this work, we aim to align and compare recommendations on evaluative reporting of DNA results. Some of these recommendations are explicitly stated while others are woven into the text. All documents pertain to evaluative reporting; they agree on the use of likelihood ratios and the need to avoid the transposed conditional. There is some disagreement such as whether a quantitative or qualitative LR should be reported. However, the majority situation is that one topic is covered explicitly by one document but only implicitly covered or not mentioned in the others. We identify 19 consensus recommendations and highlight five gaps or areas of disagreement for which we offer suggestions. It is our hope that this will encourage conversations that will lead to a more uniform set of guidelines, perhaps during a periodic updating of existing documents.

The Y chromosomal haplotype is expected to be identical (or close to, depending on the mutation rate) among a male and many of his paternal relatives. This means that often the same evidential value for the DNA evidence is obtained, whether the true donor or one of his close paternal relatives is compared to a crime sample. Commentators (see for example the UK Forensic Science Regulator or Amorim) have suggested to change the proposition pair to compare the probability of the evidence if the Person of Interest (POI) or one of his close paternal relatives left the DNA to the probability of the evidence if an unrelated male from the population left the DNA. We argue that this is problematic because there is no clear definition of close paternal relatives and truly unrelated males do not exist. Instead, we take a starting point in the traditional proposition pair "The source of the male DNA is the POI" versus "The source of the male DNA is not the POI" and make the latter one operational by suggesting that it is formulated as "The source of the male DNA is a random man from the population". The issue of matching males in the POI's lineage is then addressed either in a comment in the statement or directly through a probability model.

The subject of inter- and intra-laboratory inconsistency was recently raised in a commentary by Itiel Dror. We re-visit an inter-laboratory trial, with which some of the authors of this current discussion were associated, to diagnose the causes of any differences in the likelihood ratios (LRs) assigned using probabilistic genotyping software. Some of the variation was due to different decisions that would be made on a case-by-case basis, some due to laboratory policy and would hence differ between laboratories, and the final and smallest part was the run-to-run difference caused by the Monte Carlo aspect of the software used. However, the net variation in LRs was considerable. We believe that most laboratories will self-diagnose the cause of their difference from the majority answer and in some, but not all instances will take corrective action. An inter-laboratory exercise consisting of raw data files for relatively straightforward mixtures, such as two mixtures of three or four persons, would allow laboratories to calibrate their procedures and findings.

We describe the estimation of θ (theta) values from autosomal STR sequencing data for five metapopulations. The data were compiled from 20 publications and included 39 datasets comprising a total of 7005 samples. The estimates are suitable for use within the calculation of match probabilities in forensic casework. We also have constructed a phylogenetic tree using this data that aligns with our understanding of human evolution.

We examine 31,011 PPY23 profiles at the population, metapopulation and world levels. Most haplotypes appear only once but a few have higher counts, including a set of 23 matching profiles in Delhi, India and a set of 16 matching profiles in Burkina Faso with one additional matching American African profile. We estimate FSTvalues to be used as “theta” (θ) in match probability calculations, following the method we used in our earlier survey of autosomal STR data. Match probability estimates using FˆST or the κ method of Brenner for a previously unseen profile are similar but differ for any profile previously seen.

Journal of Forensic SciencesEarly View LETTER TO THE EDITOR Commentary on: Thompson WC. Uncertainty in probabilistic genotyping of low template DNA: a case study comparing STRmix™ and TrueAllele™. J Forensic Sci. 2023;68 (3):1049–63. doi: 10.1111/1556-4029.15225 Tim Kalafut PhD, Tim Kalafut PhD Department of Forensic Science, College of Criminal Justice, Sam Houston State University, Huntsville, Texas, USASearch for more papers by this authorJames M. Curran PhD, James M. Curran PhD Department of Statistics, University of Auckland, Auckland, New ZealandSearch for more papers by this authorMichael D. Coble PhD, Michael D. Coble PhD Department of Microbiology, Immunology, and Genetics, Center for Human Identification, University of North Texas Health Science Center, Fort Worth, Texas, USASearch for more papers by this authorJohn Buckleton DSc, Corresponding Author John Buckleton DSc [email protected] Department of Statistics, University of Auckland, Auckland, New Zealand Institute of Environmental Science and Research Limited, Auckland, New Zealand Correspondence John Buckleton, Institute of Environmental Science and Research Limited, Private Bag 92021, Auckland 1142, New Zealand. Email: [email protected]Search for more papers by this author Tim Kalafut PhD, Tim Kalafut PhD Department of Forensic Science, College of Criminal Justice, Sam Houston State University, Huntsville, Texas, USASearch for more papers by this authorJames M. Curran PhD, James M. Curran PhD Department of Statistics, University of Auckland, Auckland, New ZealandSearch for more papers by this authorMichael D. Coble PhD, Michael D. Coble PhD Department of Microbiology, Immunology, and Genetics, Center for Human Identification, University of North Texas Health Science Center, Fort Worth, Texas, USASearch for more papers by this authorJohn Buckleton DSc, Corresponding Author John Buckleton DSc [email protected] Department of Statistics, University of Auckland, Auckland, New Zealand Institute of Environmental Science and Research Limited, Auckland, New Zealand Correspondence John Buckleton, Institute of Environmental Science and Research Limited, Private Bag 92021, Auckland 1142, New Zealand. Email: [email protected]Search for more papers by this author First published: 25 October 2023 https://doi.org/10.1111/1556-4029.15405Read the full textAboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onEmailFacebookTwitterLinkedInRedditWechat REFERENCES 1Thompson WC. Uncertainty in probabilistic genotyping of low template DNA: a case study comparing STRmix™ and TrueAllele™. J Forensic Sci. 2023; 68(3): 1049–1063. https://doi.org/10.1111/1556-4029.15225 10.1111/1556-4029.15225 PubMedWeb of Science®Google Scholar 2Riman S, Iyer H, Vallone PM. Examining performance and likelihood ratios for two likelihood ratio systems using the PROVEDIt dataset. PloS One. 2021; 16(9):e0256714. https://doi.org/10.1371/journal.pone.0256714 10.1371/journal.pone.0256714 CASPubMedWeb of Science®Google Scholar 3Buckleton J, Bright J-A, Taylor D, Wivell R, Bleka Ø, Gill P, et al. Re: Riman et al. Examining performance and likelihood ratios for two likelihood ratio systems using the PROVEDIt dataset. Forensic Sci Int Genet. 2022; 59:102709. https://doi.org/10.1016/j.fsigen.2022.102709 10.1016/j.fsigen.2022.102709 CASPubMedWeb of Science®Google Scholar 4Ramos D, Meuwly D, Haraksim R, Berger CEH. Validation of forensic automatic likelihood ratio methods. In: D Banks, K Kafadar, D Kaye, M Tackett, editors. Handbook of forensic statistics. Boca Raton, Fl: Chapman & Hall/CRC; 2020. 10.1201/9780367527709-7 Google Scholar 5Buckleton JS, Kruijver M, Curran J, Bright J-A. Calibration of STRmix LRs following the method of Hannig et al. Institute of Environmental Science and Research. 2020 https://research.esr.cri.nz/articles/report/Calibration_of_STRmix_LRs_following_the_method_of_Hannig_et_al_/12324011. Accessed 05 Oct 2023 Google Scholar 6Bright J-A, Jones Dukes M, Pugh SN, Evett IW, Buckleton JS. Applying calibration to LRs produced by a DNA interpretation software. Aus J Forensic Sci. 2021; 53(2): 147–153. https://doi.org/10.1080/00450618.2019.1682668 10.1080/00450618.2019.1682668 Web of Science®Google Scholar 7Russell L, Cooper S, Wivell R, Kerr Z, Taylor D, Buckleton J, et al. A guide to results and diagnostics within a STRmix™ report. WIREs Forensic Sci. 2019; 1(6):e1354. https://doi.org/10.1002/wfs2.1354 10.1002/wfs2.1354 Google Scholar 8Bright J-A, Richards R, Kruijver M, Kelly H, McGovern C, Magee A, et al. Internal validation of STRmix™ – a multi laboratory response to PCAST. Forensic Sci Int Genet. 2018; 34: 11–24. https://doi.org/10.1016/j.fsigen.2018.01.003 10.1016/j.fsigen.2018.01.003 CASPubMedWeb of Science®Google Scholar 9Biedermann A, Taroni F, Thompson WC. Using graphical probability analysis (Bayes nets) to evaluate a conditional DNA inclusion. Law Probab Risk. 2011; 10(2): 89–121. https://doi.org/10.1093/lpr/mgr007 10.1093/lpr/mgr007 Google Scholar 10Slooten K, Caliebe A. Contributors are a nuisance (parameter) for DNA mixture evidence evaluation. Forensic Sci Int Genet. 2018; 37: 116–125. https://doi.org/10.1016/j.fsigen.2018.05.004 10.1016/j.fsigen.2018.05.004 CASPubMedWeb of Science®Google Scholar 11McGovern C, Bright J-A, Buckleton J, Taylor D. Examination of a novel method to allow for a range in the number of contributors to a DNA profile: a key area of subjectivity in DNA evidence interpretation. Proceedings of the 26th international symposium of human identification. Grapevine, TX. Madison, WI: Promega Corporation; 2015 https://scv10mr-cdnpre-p-cus-00.azureedge.net/-/media/files/products-and-services/genetic-identity/ishi-26-oral-abstracts/14-mcgovern.pdf?rev=767064bfd91b4b70b755bfbcae5f1a3f&_gl=1*1mdvymo*_ga*NjA2NTY2OS4xNjkxMzQ0Mzkx*_ga_3KTGR4JGE1*MTY5MTM0NDM5MC4xLjAuMTY5MTM0NDM5Mi42MC4wLjA. Accessed 05 Oct 2023 Google Scholar 12Taylor D, Bright J-A, Buckleton J. Interpreting forensic DNA profiling evidence without specifying the number of contributors. Forensic Sci Int Genet. 2014; 13: 269–280. https://doi.org/10.1016/j.fsigen.2014.08.014 10.1016/j.fsigen.2014.08.014 CASPubMedWeb of Science®Google Scholar 13 Scientific Working Group on DNA Analysis Methods. Recommendations of the SWGDAM Ad Hoc Working Group on genotyping results reported as likelihood ratios. 2018 https://www.swgdam.org/_files/ugd/4344b0_dd5221694d1448588dcd0937738c9e46.pdf. Accessed 05 Oct 2023 Google Scholar 14Taroni F, Bozza S, Biedermann A, Aitken C. Dismissal of the illusion of uncertainty in the assessment of a likelihood ratio. Law Probab Risk. 2015; 15(1): 1–16. https://doi.org/10.1093/lpr/mgv008 10.1093/lpr/mgv008 Web of Science®Google Scholar 15Meester R, Slooten K. Probability and forensic evidence: theory, philosophy, and applications. Cambridge, U.K.: Cambridge University Press; 2021. https://doi.org/10.1017/9781108596176 10.1017/9781108596176 Google Scholar 16Berger CEH, Slooten K. The LR does not exist. Sci Justice. 2016; 56(5): 388–391. https://doi.org/10.1016/j.scijus.2016.06.005 10.1016/j.scijus.2016.06.005 PubMedWeb of Science®Google Scholar 17Thompson W. Painting the target around the matching profile: the Texas sharpshooter fallacy in forensic DNA interpretation. Law Probab Risk. 2009; 8: 257–276. https://doi.org/10.1093/lpr/mgp013 10.1093/lpr/mgp013 Google Scholar Early ViewOnline Version of Record before inclusion in an issue ReferencesRelatedInformation

We consider both the theory and implementation of uncertainty in the likelihood ratio (LR) produced by the probabilistic genotyping software STRmix™ when analysing forensic DNA evidence.If a reasonable effort can be made to integrate out most of the identifiable uncertainty in assigning an LR, then this estimate is the best estimate to report. We are not certain we are capable of this at writing although considerable progress has been made. We recognise that the community and stakeholders are very comfortable with deliberate understatement of the LR and are not likely to embrace this concept.The lower bound produced by STRmix™ is often but not always conservative depending on the circumstances of the case. If the upper and lower bounds bracket LR = 1, then it may be best to report the comparison as inconclusive.If both upper and lower bounds are below one, then it is necessary to decide whether it is better to understate the evidence against the person of interest (POI) or understate the inference possible from these data.

One of the Daubert standard’s illustrative factors is a known error rate. Gross errors, such as contamination and sample swapping do happen, as well as many errors with lesser consequences. However, many commentators seek to include false support for non-donors under the classification of error. This appears to be a terminology difference between scientists and lawyers. Scientists would not consider these errors but rather as part of the expected performance of any system with less than 100% discrimination.Using what we presume is the legal view that false support is an error, again some commentators seek to inform the rate of this using the exceedance probability or the p-value. This is the probability that a false donor would give an LR greater than 1 or greater than the LR in this case, respectively. This erroneously treats the LR as a categorical variable. The exceedance probability systematically underrepresents the weight of evidence, and the p-value can overstate the value of the evidence. The only route to the correct logic is to combine the LR and the prior odds.

Thompson reports a comparison of data from STRmix and TrueAllele. The data he has arises from different inputs to the two software. If the input data are made more similar the outputs become more similar. Thompson argues that the Analytical Threshold, AT, should be varied in casework. This produced different LRs but the analyst would be left deciding what to do with these options. This cannot be based on the LRs but should be based on whether any movement in the AT adds reliable or unreliable data. This is how most laboratories set their AT in the first place. Hence it is pointless, and potentially dangerous, to experimentally vary the AT in casework. The profile is low level and shows at most three peaks. Thompson argues that LR results assuming that the number of contributors (NoC) is 2 or 3 should be reported. Uncertainty in NoC should be treated as a nuisance variable and summed out.

Simple propositions are defined as those with one POI and the remaining contributors unknown under Hp and all unknown contributors under Ha. Conditional propositions are defined as those with one POI, one or more assumed contributors, and the remaining contributors (if any) unknown under Hp, and the assumed contributor(s) and N unknown contributors under Ha. In this study, compound propositions are those with multiple POI and the remaining contributors unknown under Hp and all unknown contributors under Ha. We study the performance of these three proposition sets on thirty-two samples (two laboratories × four NOCs × four mixtures) consisting of four mixtures, each with N = 2, N = 3, N = 4, and N = 5 contributors using the probabilistic genotyping software, STRmix™. In this study, it was found that conditional propositions have a much higher ability to differentiate true from false donors than simple propositions. Compound propositions can misstate the weight of evidence given the propositions strongly in either direction.

Complexity thresholds have previously been proposed as a rejection criterion for when a forensic DNA profile is potentially too complicated for analysis. Queensland Health Forensic and Scientific Services (FSS) had implemented a complexity threshold between an assigned number of contributors (NoC) of three and four. This was based on having validated their software for NoC 1 to 3 and only specific four person mixtures. Results were reported in real time as they came individually off the process. This resulted in initial reports that indicated strong inclusionary evidence that may have initiated Police and court action. These initial reports could be withdrawn at a later stage when the NoC was revised from three to four. This situation was exacerbated by a systematic tendency at FSS to overstate NoC.Data presented in this paper reconfirms the previously known observation that a donor that gives a high LR at NoC = 3 will also give a high LR at NoC = 4. This is an example of the potential unintended consequences of the use of hard thresholds such as complexity thresholds. It signals that care is needed when making sensible criteria for implementation in casework.

There is interest in comparing the output, principally the likelihood ratio, from the two probabilistic genotyping software EuroForMix (EFM) and STRmix™. Many of these comparison studies are descriptive and make little or no effort to diagnose the cause of difference. There are fundamental model differences between EFM and STRmix™ that are causative of any likelihood ratio differences. These model differences are easily identified, and it is certainly possible to arbitrate on whether the models in none, one, or both software are the more supported. There is little point in highlighting the output differences without any effort to identify the cause as this can trigger unsubstantiated concern in those cases where a different answer is obtained. In this paper, we further analyse the models within EFM and STRmix™ and explain their differences.

There is interest in comparing the output, principally the likelihood ratio, from the two probabilistic genotyping software EuroForMix (EFM) and STRmix™. Many of these comparison studies are descriptive and make little or no effort to diagnose the cause of difference. There are fundamental differences between EFM and STRmix™ that are causative of the largest set of likelihood ratio differences. This set of differences is for false donors where there are many instances of LRs just above or below 1 for EFM that give much lower LRs in STRmix™. This is caused by the separate estimation of parameters such as allele height variance and mixture proportion using MLE under Hp and Ha for EFM. This can result in very different estimations of these parameters under Hp and Ha . It results in a departure from calibration for EFM in the region of LRs just above and below 1.

Evidential value of DNA mixtures is typically expressed by a likelihood ratio. However, selecting appropriate propositions can be contentious, because assumptions may need to be made around, for example, the contribution of a complainant's profile, or relatedness between contributors. A choice made one way or another disregards any uncertainty that may be present about such an assumption. To address this, a complex proposition that considers multiple sub-propositions with different assumptions may be more appropriate. While the use of complex propositions has been advocated in the literature, the uptake in casework has been limited. We provide a mathematical framework for evaluating DNA evidence given complex propositions and discuss its implementation in the DBLR™ software. The software simultaneously handles multiple mixed samples, reference profiles and relationships as described by a pedigree, which unlocks a variety of applications. We provide several examples to illustrate how complex propositions can efficiently evaluate DNA evidence. The addition of this feature to DBLR™ provides a tool to approach the long-accepted, but often impractical suggestion that propositions should be exhaustive within a case context.

Standard processing of electrophoretic data within a forensic DNA laboratory is for one (or two) analysts to designate peaks as either artefactual or non-artefactual in a process commonly referred to as profile 'reading'. Recently, FaSTR™ DNA has been developed to use artificial neural networks to automatically classify fluorescence within an electropherogram as baseline, allele, stutter or pull-up. These classifications are based on probabilities assigned to each timepoint (scan) within the electropherogram. Instead of using the probabilities to assign fluorescence into a category they can be used directly in the profile analysis. This has a number of advantages; increased objectivity in DNA profile processing, the removal for the need for analysts to read profiles, the removal for the need of an analytical threshold. Models within STRmix™ were extended to incorporate the peak label probabilities assigned by FaSTR™ DNA. The performance of the model extensions was tested on a DNA mixture dataset, comprising 2-4 person samples. This dataset was processed in a 'standard' manner using an analytical threshold of 50rfu, analyst peak designations and STRmix™ V2.9 models. The same dataset was then processed in an automated manner using no analytical threshold, no analysts reading the profile and using the STRmix™ models extended to incorporate peak label probabilities. Both datasets were compared to the known DNA donors and a set of non-donors. The result between the two processes was a very close performance, but with a large efficiency gain in the 0rfu process. Utilising peak label probabilities opens up the possibility for a range of workflow process efficiency gains, but beyond this allows full use of all data within an electropherogram.

We describe the developmental validation of the probabilistic genotyping software - STRmix™ NGS - developed for the interpretation of forensic DNA profiles containing autosomal STRs generated using next generation sequencing (NGS) also known as massively parallel sequencing (MPS) technologies. Developmental validation was carried out in accordance with the Scientific Working Group on DNA Analysis Methods (SWGDAM) Guidelines for the Validation of Probabilistic Genotyping Systems and the International Society for Forensic Genetics (ISFG) recommendations and included sensitivity and specificity testing, accuracy, precision, and the interpretation of case-types samples. The results of developmental validation demonstrate the appropriateness of the software for the interpretation of profiles developed using NGS technology.