The use of hair evidence for human identification is undergoing considerable improvement through the adoption of proteomic genotyping. Unlike traditional microscopic comparisons, protein sequencing provides quantitative and empirically based estimates for random match probability. Non-synonymous SNPs are translated as single amino acid polymorphisms and result in genetically variant peptides. Using high resolution mass spectrometry, these peptides can be detected in hair shaft proteins and used to infer the genotypes of corresponding SNP alleles. We describe experiments to optimize the proteomic genotyping approach to individual identification from a single human scalp hair 2 cm in length (∼100 μg). This is a necessary step to develop a protocol that will be useful to forensic investigators. To increase peptide yield from hair, and to maximize genetically variant peptide and ancestral information, we examined the conditions for reduction, alkylation, and protein digestion that specifically address the distinctive chemistry of the hair shaft. Results indicate that optimal conditions for proteomic analysis of a single human hair include 6 h of reduction with 100 mM dithiothreitol at room temperature, alkylation with 200 mM iodoacetamide for 45 min, and 6 h of digestion with two 1:50 (enzyme:protein) additions of stabilized trypsin at room temperature, with stirring incorporated into all three steps. Our final conditions using optimized temperatures and incubation times increased the average number of genetically variant peptides from 20 ± 5 to 73 ± 5 (p = 1 × 10-13), excluding intractable hair samples. Random match probabilities reached up to 1 in 620 million from a single hair with a median value of 1 in 1.1 million, compared to a maximum random match probability of 1 in 1380 and a median value of 1 in 24 for the original hair protein extraction method. Ancestral information was also present in the data. While the number of genetically variant peptides detected were equivalent for both European and African subjects, the estimated random match probabilities for inferred genotypes of European subjects were considerably smaller in African reference populations and vice versa, resulting in a difference in likelihood ratios of 6.8 orders of magnitude. This research will assure uniformity in results across different biogeographic backgrounds and enhance the use of novel peptide analysis in forensic science by helping to optimize genetically variant peptide yields and discovery. This work also introduces two algorithms, GVP Finder and GVP Scout, which facilitate searches, calculate random match probabilities, and aid in discovery of genetically variant peptides.
