Comparison of the performance of different models for the interpretation of low level mixed DNA profiles

Abstract

DNA analyses from forensic casework samples commonly result in complex DNA profiles. Often, these profiles consist of multiple contributors and display multiple stochastic events such as peak height imbalance, allelic or locus drop-out, allelic drop-in, and excessive or indistinguishable stutter. This increased complexity has established a need for more sophisticated methods of DNA mixture interpretation. This study compares the effectiveness of statistical models in the interpretation of artificially created low template two person mixed DNA profiles at varying proportions and template quantities. Two binary models (combined probability of inclusion and random match probability), a semicontinuous (Lab retriever), and continuous model (STRmix™) were compared. Generally, as the sophistication of the models increases, the power of discrimination increases. Differences in discrimination often correlate to each model's ability to use observed data effectively. Binary models require static thresholds resulting in unused data and outliers that may lead to difficult or incorrect interpretation. Semicontinuous and continuous models eliminate the stochastic threshold, however Lab Retriever does not account for stochastic events beyond drop-out and drop-in leading to possible less effective use of the data. STRmix™ incorporates all stochastic events listed above into the calculation making the most effective use of the observed data.

Evaluation of samples comprising minute amounts of DNA

“…From our results (that used a discrete/semi-continuous LR model), we conclude that when the relevance of the evidence is weak, it may be best to decide not to continue to computing an LR, even though it could result in a large LR in favour of the prosecution hypothesis. The specifics of the LR-based statistical model (such as the choice between a continuous and semi-continuous model that make yes or no use of peak height information) that is applied, will affect the actual LR that is obtained [23,[35][36][37][38][39]. The suitability of using peak height information in profiles derived with enhanced interrogation techniques has, however, not been shown.…”

Probabilistic peak detection in CE‐LIF for STR DNA typing

“…These peaks were identified by the Bayesian approach but excluded by the “GeneMapperID” software when an AT = 50 RFU was applied. Though several recent studies have reported the use of, what is colloquially termed, subthreshold data and have shown improved performance, we present a novel probabilistic method that is not dependent upon determinations of peak presence; rather, we present a probabilistic peak detection algorithm that may, in future, be intergraded in the DNA pipeline.…”

“…The continuous models make use of the peak height information and thus utilize an additional source of information from the epg. These approaches include the modeling of stutter phenomena, different number of contributors, and allele drop‐in and drop‐out, and the forensic community has in recent years introduced such methods and associated software solutions (both open source and commercial) for probabilistic evaluation and interpretation . To enable the statistical evaluation of epgs with alleles under the threshold, these software tools can assess any genotype candidate value to determine how well it explains the observed data.…”

An alternative application of the consensus method to DNA typing interpretation for Low Template-DNA mixtures

“…These models present different degrees of complexity in terms of application and comprehensibility. The most comprehensible model is the binary one [3]. However, it is broadly accepted that binary models are not suitable for complex mixtures and LT DNA evaluations, since they do not take into account several important parameters and stochastic effects such as drop-out and drop-in and, above all, thresholds (in terms of limit of detection/analytical threshold and limit of quantitation) and peak heights data.…”