Quantifying the Accuracy of Two Innovative Forensic Genetic Identification Techniques: Genealogical Searching and Low-Template DNA Mixture Analysis

Two novel techniques have dramatically expanded the power of forensic identification in recent years. First, genealogical searches allow investigators to use DNA profiles to identify distant genetic relatives of unknown individuals, helping them identify suspects. Second, low-template DNA mixture analysis allows investigators to analyze miniscule amounts of biological material, often from multiple donors. Both techniques are being rapidly deployed by investigators across the nation. Past research focused on how the accuracy of these methods is impacted by technical factors, like the number of relatives or allelic drop-out. However, less attention has been paid to population genetic factors. Therefore, our objective is to rigorously test the accuracy of both novel methods in realistic and varied population genetic scenarios. For instance, in genealogical searching, if a cryptically distant genetic relative shares a segment of DNA with a suspect, investigators may target the family of that relative, while missing the family of the offender. Similarly, when analyzing a low-template DNA mixture, if the frequency of genetic variants in a population is inappropriately estimated, a suspect may appear to have contributed to the sample, even though he did not. Thus, investigators must understand the population genetics-related limitations of the techniques they are using, so they can accurately interpret results. To this end, we will use computer simulations and existing datasets to quantify the methods’ performance under various circumstances. We will 1) Characterize a freely available pedigree dataset that forensic investigators can use to conduct research on genealogical searches. We will quantify missing data in a very large empirical pedigree dataset, and develop software to simulate similar complete pedigrees. 2) Quantify the accuracy of genealogical genetic searching, particularly the impact of shared IBD between cryptically distant relatives. We will use analytical and simulation approaches to quantify IBD sharing on a realistic pedigree, and use this information to estimate the precision and power of genealogical searching. 3) Quantify the accuracy of low-template DNA analysis to identify mixture contributors with varying population genetic backgrounds. We will leverage global allele frequency data to quantify how contributor identification accuracy varies for individuals of different genetic ancestries, particularly when ancestry is misspecified. This work will give context for the criminal justice community to better interpret results obtained from both emerging forensic genetic techniques. In addition, our work will identify the limitations of each method, which can subsequently be addressed by forensic scientists, thus reducing errors in forensic genetic identification.

Note: This project contains a research and/or development component, as defined in applicable law, and complies with Part 200 Uniform Requirements - 2 CFR 200.210(a)(14).

CA/NCF