Extending the discussion on inconsistency in forensic decisions and results.

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.

An inter-laboratory comparison of probabilistic genotyping parameters and evaluation of performance on DNA mixtures from different laboratories.

Probabilistic genotyping (PG) is becoming the preferred standard for evidence interpretation, amongst forensic DNA laboratories, especially those in the United States. Various groups have expressed concern about reliability of PG systems, especially for mixtures beyond two contributors. Studies involving interlaboratory testing of known mixtures have been identified as ways to evaluate the reliability of PG systems. Reliability means different things in different contexts. However, it suffices here to think about it as a mixture of precision and accuracy. We might also consider whether a system is prone to producing misleading results - for example large likelihood ratios (LRs) when the POI is truly not a contributor, or small LRs when the POI is a truly a contributor. In this paper we show that the PG system STRmix™ is relatively unaffected by differences in parameter settings. That is, a DNA mixture that is analyzed in different laboratories using STRmix™ will result in different LRs, but less than 0.05% of these LRs would result in a different, or misleading conclusion as long as the LR is greater than 50. For the purposes of this study, we define LRs assigned using different parameters for the same mixtures as similar if the LR of the true POI is greater than the LRs generated for 99.9% of the general population. These findings are based on an interlaboratory study involving eight laboratories that provided twenty known DNA mixtures of two to four contributors and their individual laboratory STRmix™ parameters. The eight sets of laboratory parameters included differences in STR kits and PCR cycles as well as the peak, stutter, and locus specific amplification efficiency variances.

Guiding proposition setting in forensic DNA interpretation.

There is a general reluctance to use conditioning profiles when forming propositions for cases where the evidence is a DNA mixture. However, the use of conditioning profiles improves the ability to differentiate true from false donors. There are at least four situations where this decision making is at its most difficult. These are:Rigorous mathematical treatment, given by Slooten and others, appears to offer strong guidance for these situations. This treatment assumes that the prior probabilities for conditioning, or not conditioning, on any individual are not extreme. It is when these prior probabilities appear ambiguous that the decision to condition or not can appear to be problematic. This is often the situation found in casework. In this paper we attempt to show that such situations may benefit most from following such guidance. A lower bound on the Bayes factor can be obtained by finding the highest LR that includes the POI and dividing by the highest LR that does not include the POI. These two highest LRs may be found with and without the disputed conditioning profile. The resultant lower bound is on the BF for the inclusion of the POI without directly assuming the disputed conditioning profile. Adopting this approach would both minimize adventitious inclusions and approximate an exhaustive set of propositions.

Detection and analysis of DNA mixtures with the MiSeq FGx®.

Recognizing and interpreting mixtures are challenges that occur frequently in forensic casework. Therefore, any new analysis methods that are implemented must handle the challenges of mixed forensic samples. Next generation sequencing offers advantages over capillary electrophoresis in amplicon multiplexing and degraded sample analysis; however, advantages with mixed samples rely heavily on the advancement of user-friendly analysis software. This research analyzed samples with the ForenSeq™ DNA Signature Prep Kit on the MiSeq FGx® and compared them with the GlobalFiler™ STR Kit for capillary electrophoresis. Metrics tested for both chemistries included concordance, limits of detection, and mixture analysis. Data analysis for mixture samples was completed with the MixtureAce™ plug-in and ArmedXpert™ software. Next generation sequencing offered distinct advantages in limits of detection and isoallele heterozygosity but suffered from increased variability in stutter and allele count ratios compared to capillary electrophoresis.

Study of CTS DNA Proficiency Tests with Regard to DNA Mixture Interpretation: A NIST Scientific Foundation Review.

The National Institute of Standards and Technology has released a document entitled DNA Mixture Interpretation: A NIST Scientific Foundation Review for public comment. This has become known as the Draft NIST Foundation Review. It contains the statement: "Across these 69 data sets, there were 80 false negatives and 18 false positives reported from 110,408 possible responses (27,602 participants × two evidence items × two reference items). In the past five years, the number of participants using PGS has grown." We examine a set of proficiency test results to determine if these NIST statements could be justified. The summary reports for each relevant forensic biology test (Forensic Biology, Semen, and Mixture) in the years 2018-2021 were reviewed. Data were also provided to us by CTS upon our request. None of the false positives or negatives could be attributed to the mixture interpretation strategy and certainly not to the use of PGS.

Investigation into the effect of mixtures comprising related people on non-donor likelihood ratios, and potential practises to mitigate providing misleading opinions.

The interpretation of mixtures containing related individuals can be difficult due to allele sharing between the contributors. Challenges include the assignment of the number of contributors (NoC) to the mixture with the under assignment of NoC resulting in false exclusions of true donors. Non-donating relatives of the true contributors to mixtures of close relatives can result in likelihood ratios supporting their adventitious inclusion within the mixture. We examine the effect of non-donor likelihood ratios on mixtures of first order relatives. Mixtures of full siblings and parent-child were created by mixing the DNA from known family members in vitro, or by in silico simulation. Mixtures were interpreted using the probabilistic genotyping software STRmix™ and likelihood ratios were assigned for the true donors and non-donors who were either further relatives of the true donors or unrelated to the true donors. The two donor balanced mixtures deconvoluted straightforwardly when analysed as NoC = 2 giving approximately the experimental design 1:1 ratio. When analysed as NoC = 3 a very large number of non-donor genotypes produced LRs close to 1 including many instances of adventitious support. The in vitro three donor balanced mixtures proved difficult to assign as NoC = 3 by a blind examination of the profile. It is likely that many of these would be misassigned as NoC = 2. The analysis of the in vitro and in silico mixtures assuming NoC = 3 with no use of a conditioning profile or with the use of a conditioning profile but without informed priors on the mixture proportions (Mx priors) was ineffective. If the profile can be assigned as NoC = 3 then assignment of the Mx priors is straightforward. This analysis gave no false exclusions. Adventitious support did happen for relatives with high allele sharing. Adventitious support was not observed for any unrelated non-donors. The analysis of the three-person mixtures as NoC = 2 produced many false exclusions and fewer instances of adventitious support. The three donor unbalanced mixtures could all be assigned as NoC= 3. Analysis without Mx priors produced an alternate genotype explanation.

A mixed DNA profile controversy revisited.

Semaan et al. (J Forensic Res, 2020, 11, 453) discuss a mock case "where eight different individuals [P1 through P8 ] could not be excluded in a mixed DNA analysis. Even though … expert DNA mixture analysis software was used." Two of these are the true donors. The LRs reported are incorrect due to the incorrect entry of propositions into LRmix Studio. This forced the software to account for most of the alleles as drop-in, resulting in LRs 60-70 orders of magnitude larger than expected. P1 , P2 , P4 , P5 , and P8 can be manually excluded using peak heights. This has relevance when using LRmix which does not use peak heights. We extend the work using the same two reference genotypes who were the true contributors as Semaan et al. (J Forensic Res, 2020, 11, 453). We simulate three two-donor mixtures with peak heights using these two genotypes and analyze using STRmix™. For the simulated 1:1 mixture, one of the non-donors' LRs supported him being a contributor when no conditioning was used. When considered in combination with any other potential donors (i.e., with conditioning), this non-donor was correctly eliminated. For the 3:1 mixture, all results correctly supported that the non-donors were not contributors. The low-template 4:1 mixture LRs with no conditioning showed support for all eight profiles as donors. However, the results from pair-wise conditioning showed that only the two ground truth donors had LRs supporting that they were contributors to the mixture. We recommend the use of peak heights and conditioning profiles, as this allows better sensitivity and specificity even when the persons share many alleles.

A review of likelihood ratios in forensic science based on a critique of Stiffelman "No longer the Gold standard: Probabilistic genotyping is changing the nature of DNA evidence in criminal trials".

Stiffelman [1] gives a broad critique of the application of likelihood ratios (LRs) in forensic science, in particular their use in probabilistic genotyping (PG) software. These are discussed in this review. LRs do not infringe on the ultimate issue. The Bayesian paradigm clearly separates the role of the scientist from that of the decision makers and distances the scientist from comment on the ultimate and subsidiary issues. LRs do not affect the reasonable doubt standard. Fact finders must still make decisions based on all the evidence and they must do this considering all evidence, not just that given probabilistically. LRs do not infringe on the presumption of innocence. The presumption of innocence does not equate with a prior probability of zero but simply that the person of interest (POI) is no more likely than anyone else to be the donor. Propositions need to be exhaustive within the context of the case. That is, propositions deemed relevant by either defense or prosecution which are not fanciful must not be omitted from consideration.

Using the Nondonor Distribution to Improve Communication and Inform Decision Making for Low LRs from Minor Contributors in Mixed DNA Profiles.

The reporting of a likelihood ratio (LR) calculated from probabilistic genotyping software has become more popular since 2015 and has allowed for the use of more complex mixtures at court. The meaning of "inconclusive" LRs and how to communicate the significance of low LRs at court is now important. We present a method here using the distribution of LRs obtained from nondonors. The nondonor distribution is useful for examining calibration and discrimination for profiles that have produced LRs less than about 104 . In this paper, a range of mixed DNA profiles of varying quantity were constructed and the LR distribution considering the minor contributor for a number of nondonors was compared to the expectation given a calibrated system. It is demonstrated that conditioning genotypes should be used where reasonable given the background information to decrease the rate of nondonor LRs above 1. In all 17 cases examined, the LR for the minor donor was higher than the nondonor LRs, and in 12 of the 17 cases, the 99.9 percentile of the nondonor distribution was lower when appropriate conditioning information was used. The output of the tool is a graph that can show the position of the LR for the person of interest set against the nondonor LR distribution. This may assist communication between scientists and the court.

Implementation and validation of an improved allele specific stutter filtering method for electropherogram interpretation.

Modern probabilistic genotyping (PG) software is capable of modeling stutter as part of the profile weighting statistic. This allows for peaks in stutter positions to be considered as allelic or stutter or both. However, prior to running any sample through a PG calculator, the examiner must first interpret the sample, considering such things as artifacts and number of contributors (NOC or N). Stutter can play a major role both during the assignment of the number of contributors, and the assessment of inclusion and exclusion. If stutter peaks are not filtered when they should be, it can lead to the assignment of an additional contributor, causing N contributors to be assigned as N + 1. If peaks in the stutter position of a major contributor are filtered using a threshold that is too high, true alleles of minor contributors can be lost. Until now, the software used to view the electropherogram stutter filters are based on a locus specific model. Combined stutter peaks occur when a peak could be the result of both back stutter (stutter one repeat shorter than the allele) and forward stutter (stutter one repeat unit larger than the allele). This can challenge existing filters. We present here a novel stutter filter model in the ArmedXpert™ software package that uses a linear model based on allele for back stutter and applies an additive filter for combined stutter. We term this the allele specific stutter model (AM). We compared AM with a traditional model based on locus specific stutter filters (termed LM). This improved stutter model has the benefit of: Instances of over filtering were reduced 78% from 101 for a traditional model (LM) to 22 for the allele specific model (AM) when scored against each other. Instances of under filtering were reduced 80% from 85 (LM) to 17 (AM) when scored against ground truth mixtures.

Internal validation of STRmix™ - A multi laboratory response to PCAST.

We report a large compilation of the internal validations of the probabilistic genotyping software STRmix™. Thirty one laboratories contributed data resulting in 2825 mixtures comprising three to six donors and a wide range of multiplex, equipment, mixture proportions and templates. Previously reported trends in the LR were confirmed including less discriminatory LRs occurring both for donors and non-donors at low template (for the donor in question) and at high contributor number. We were unable to isolate an effect of allelic sharing. Any apparent effect appears to be largely confounded with increased contributor number.

A Practical Guide for the Formulation of Propositions in the Bayesian Approach to DNA Evidence Interpretation in an Adversarial Environment.

The interpretation of complex DNA profiles is facilitated by a Bayesian approach. This approach requires the development of a pair of propositions: one aligned to the prosecution case and one to the defense case. This note explores the issue of proposition setting in an adversarial environment by a series of examples. A set of guidelines generalize how to formulate propositions when there is a single person of interest and when there are multiple individuals of interest. Additional explanations cover how to handle multiple defense propositions, relatives, and the transition from subsource level to activity level propositions. The propositions depend on case information and the allegations of each of the parties. The prosecution proposition is usually known. The authors suggest that a sensible proposition is selected for the defense that is consistent with their stance, if available, and consistent with a realistic defense if their position is not known.

Validating multiplexes for use in conjunction with modern interpretation strategies.

In response to requests from the forensic community, commercial companies are generating larger, more sensitive, and more discriminating STR multiplexes. These multiplexes are now applied to a wider range of samples including complex multi-person mixtures. In parallel there is an overdue reappraisal of profile interpretation methodology. Aspects of this reappraisal include 1. The need for a quantitative understanding of allele and stutter peak heights and their variability, 2. An interest in reassessing the utility of smaller peaks below the often used analytical threshold, 3. A need to understand not just the occurrence of peak drop-in but also the height distribution of such peaks, and 4. A need to understand the limitations of the multiplex-interpretation strategy pair implemented. In this work we present a full scheme for validation of a new multiplex that is suitable for informing modern interpretation practice. We predominantly use GlobalFiler™ as an example multiplex but we suggest that the aspects investigated here are fundamental to introducing any multiplex in the modern interpretation environment.