326V Poster Online - Virtual Posters
Tuesday June 07, 11:00 AM - 3:00 PM

Authors:
Cara Ly ¹; Kamillah Felix ¹; Evan Ho ¹; Chris Godek ¹; Niquo Ceberio ¹; Maria Flores ²; Hannah Thorner ³; Carina Kalaydjian ²; Matt Paunovich ¹; Rori Rohlfs ¹

Affiliations:
1) San Francisco State University; 2) University of California, Los Angeles; 3) George Washington University

Keywords:
Other (Forensics)

Police are increasingly using trace amounts of DNA in investigations, often when samples contain DNA from multiple contributors. The reliability of interpreting these DNA mixtures is complex. A likelihood ratio (LR) is a value used to assess whether genetic evidence at the crime scene supports the defense hypothesis (the suspect did not contribute to the DNA mixture), or the prosecutor's hypothesis (the suspect did contribute to the mixture). The LR is calculated based on the allele frequency distribution of the assumed population group that the suspect belongs to, which will be referred to as the reference population. When the reference population's allele frequency is different from the suspect's true genetic background, the accuracy of the LR results can be impacted. We hypothesized that there will be higher error rates when the reference population is incorrectly assumed, and even higher when the true population has low genetic diversity. We also expected to see that as the number of contributors in a mixture increases, the amount of false positives increases. To test our hypothesis, we used Forensim – a free open source R package – to simulate individual genotypes, forensic DNA mixtures, and calculate the LR, which allows us to compare the probability of the data for when the suspect does and does not contribute to the mixture. We observed that when there are more individuals in the DNA mixture, there is less reliability in assessing the evidence that a suspect did not contribute. With the correct reference population, there was an increase in the false positive rate for populations with low genetic diversity compared to populations with high genetic diversity. We identified more false positives with an incorrect reference population and an even higher false positive rate when the true population of the suspect has low genetic diversity and the reference population has much higher genetic diversity. Our results indicate that forensic DNA mixture analysis tools used today may falsely identify individuals with certain genetic backgrounds due to differences in both genetic diversity and the appropriateness of reference populations.