Establishing STR and identity SNP analysis thresholds for reliable interpretation and practical implementation of MPS gDNA casework

Abstract

Massively parallel sequencing (MPS) evaluation by the global forensic community has demonstrated immediate advantages and forthcoming, versatile capabilities to aid missing persons and criminal investigations. As a result, an increased number of operational casework laboratories are executing validation geared towards informing MPS interpretation guidelines and SOPs for routine implementation. As with CE-based data, establishing analysis thresholds that enable reliable interpretation within the limitations of a system is a pivotal decision point. In this article, we present an empirical data-driven approach for determining analytical and stochastic thresholds (AT, ST) for the 152 STR (auto, Y, and X) and identity-informative SNP (iiSNP) loci targeted in the ForenSeq™ DNA Signature assay. While the methods utilized mimic CE threshold determination, fundamental distinctions specific to MPS data are addressed. Key considerations include relative and fixed read count values as well as interlocus balance expectations when increasing numbers of markers and/or marker types are multiplexed together in a single amplification. Internal data as well as data from collaborating laboratories will be presented. For AT assessment, sensitivity data (1 ng to 6 pg gDNA input) were evaluated with a combination of percent-based and static minimum read count AT values. Comparison of varying AT values, their effect on true allele call rates, and unanticipated read detection were assessed to achieve an ideal balance of data quality and quantity. Additionally, 4 theoretical options for ST application will be offered which include a detailed review of the stepwise methods, calculations, and resulting data that support a locus “grouping” approach based on locus performance, or intensity groups, within each STR and iiSNP marker set.