This study explores computational methods for the interpretation of forensic DNA samples.
This study aims to develop computational methods for the analysis of STR profiles that are robust to these phenomena and that utilize quantitative peak height information captured in profiles. These methods are expected to improve significantly on existing methods for analysis of STR profiles, particularly in cases of low amounts of template DNA or where there are many contributors. The researchers began by characterizing the distribution of signal, noise, and stutter peak heights and studied their dependence on template DNA amount. For the second part of the project, the researchers developed a method to identify the number of contributors to a DNA sample. The researchers’ method, NOCIt, calculates the a posteriori probability on the number of contributors to a forensic sample, taking into account signal peak heights, population allele frequencies, baseline noise, allele dropout, and stutter. On the experimental samples tested, NOCIt had an accuracy of 83%, while the accuracy of the best pre-existing method was 72%. The accuracies of NOCIt and the best pre-existing method on the simulated profiles were 85% and 73%. The researchers were able to reduce the running time of NOCIt by developing a faster method based on an importance sampling algorithm. Finally, the researchers developed a computational tool (MatchIt) to directly compute a continuous Likelihood Ratio (LR) for a person of interest (POI), treating other contributors (if any) as interference. MatchIt also calculates the distribution of the LR along with the p-value, which is the probability a randomly chosen individual results in a LR at least as large as the LR obtained from the POI. The researchers observed that the amount of template DNA from the contributor impacted the LR – small LRs resulted from contributors with low template masses. Moreover, the researchers observed a decrease of p-values as the LR increased.