Robust detection of individual forensic profiles in DNA mixtures

Highlights • A panel of 480 low minor allele frequency SNPs was developed to reliably detect contributors in complex DNA mixtures. • DNA from 3, 5, 8, 10 and 15-person mixtures was sequenced and analyzed. • A statistical model of Probability of Random Man Not Excluded [ P ( RMNE )] was developed. P( RMNE ) model incorporates varying allele frequency across loci and mismatched alleles in the mixture and reference profiles. • With zero mismatches, detection of individuals can be obtained in 10-person mixtures equimolar DNA samples with a P ( RMNE ) of 5.9 × 10−6. • The method is being extended to larger SNP panels and forensic “touch” samples.