Four Model Variants Within a Continuous Forensic DNA Mixture Interpretation Framework: Effects on Evidential Inference and Reporting

• NCJ Number: 252470
• Author(s): Harish Swaminathan, Muhammad O. Qureshi, Catherine M. Grgicak, Ken Duffy, Desmond S. Lun
• Date Published: November 2018
• Length: 23 pages
• Agencies: NIJ-Sponsored
• Award: Establishing Exclusion Criteria and the Significance of Inclusion: Developing Tools for the Interpretation of Complex DNA Mixtures; Computational methods for the interpretation of forensic DNA samples

Annotation

This project studied the impact of probabilistic model differences in DNA genotyping on the likelihood ratio (LR) in a match and on the corresponding verbal expression computed using four variants of a continuous mixture interpretation method.

Abstract

Four probabilistic models were tested five times each on 101, 1-, 2-, and 3-person experimental samples with known contributors. For each sample LRs were computed using the known contributor as the person of interest. In all four models, intra-model variability increased with an increase in the number of contributors and with a decrease in the contributor’s template mass. Inter-model variability in the associated verbal expression of the LR was observed in 32 of the 195 LRs used for comparison. In 11 of these profiles, there was a change from LR > 1 to LR < 1. These findings indicate that modifications to existing continuous models have the potential to impact significantly the final statistic. These study findings have implications for the use of and communications associated with probabilistic genotyping systems. As forensic laboratories implement probabilistic genotyping systems, characterizing the sensitivity of the LR to model assumptions of a continuous mixture interpretation method is necessary. This suggests that any updated version of existing mixture interpretation software be tested on a large number of known samples to establish the range in which the system is deemed to be reliable and to verify that its results conform to expectations. If the software is intended to be applied to low template samples, performing validation studies on such samples would inform the analyst as to the LRs typically obtained for such samples. 7 figures, 5 tables, and 46 references