Protein is a major component of all biological evidence. Proteomic genotyping is the use of genetically variant peptides that contain single amino acid polymorphisms to infer the genotype of matching non-synonymous single nucleotide polymorphisms for the individual who originated the protein sample. This can be used to statistically associate an individual to forensic evidence. The utility of the inferred genotype increases as the detection of genetically variant peptides increases and as the technology is transferred to mass spectrometry platforms available to forensic practitioners. Digests of single (2 cm) human hair shafts from three European and two African subjects were analyzed using data dependent acquisition on a Q-Exactive™ Plus Hybrid Quadrupole-Orbitrap™ system, data independent acquisition and a variant of parallel reaction monitoring on a Orbitrap Fusion™ Lumos™ Tribrid™ system, and multiple reaction monitoring on an Agilent 6495 triple quadrupole system. In our hands, average genetically variant peptide detection from a selected 24 genetically variant peptide panel increased from 6.5 ± 1.1 and 3.1 ± 0.8 using data dependent and independent acquisition to 9.5 ± 0.7 and 11.7 ± 1.7 using parallel reaction and multiple reaction monitoring (p < 0.05). Targeted methods of analysis resulted in a 1.3-fold to 1.6-fold increase in detection sensitivity. This increase in biomarker detection has a functional impact on the statistical association of a protein sample and an individual. Increased biomarker sensitivity, using Markov Chain Monte Carlo modeling, produced a median estimated random match probability of over 1 in 10 billion to 1 in 10 trillion from a single hair using targeted proteomics. Detected genetically variant peptides were validated by the inclusion of stable isotope labeled peptides in each sample, which served also as a detection trigger. This research accomplishes two aims: the demonstration of utility for alternative analytical platforms in proteomic genotyping, and the establishment of validation methods for the evaluation of inferred genotypes by forensic practitioners.

Hair shafts were collected from 5 individuals who are representative of two populous ancestral backgrounds in of the United States: European and African. The number of individuals needed for this study was minimal since this was a novel proof of concept study to demonstrate the usage of targeted proteomics for proteomic genotyping. Enough donors were used to assess reproducibility and calculate standard deviation. Three single hairs from each individual, 2 cm in length, were processed separately using a previously developed method, with a total of 15 hair digests. A blank with trypsin and without trypsin were also processed in parallel with all other digests. 24 stable isotope labeled (SIL) standard genetically variant peptides (GVPs) were spiked into the hair digests for the two targeted acquisition runs (Table S1). Raw mass spectral data were processed using the Skyline software for the 24 GVPs of interest and their corresponding heavy-isotope peptide standards. 24 peptides were chosen to adequately represent the diversity of the full set of 408 currently-identified GVPs in terms of detection sensitivity, length, and composition. Basic statistical analyses were conducted such as standard deviation, random match probability, false discovery rate (FP/(FP+TP)), and detection sensitivity (TP/(TP+FN)) to compare the three analytical methodologies. Random match probability calculations were estimated using the procedure outlined in Parker et al.

Samples used in this study were prepared as part of an earlier study. Briefly, five individuals were analyzed: three subjects of European (Davis, CA) and two subjects of African genetic background, respectively (Sorenson Forensics LLC, Salt Lake City, UT). Hair and saliva were collected using protocols compliant with the Institutional Review Board at the University of California, Davis (IRB# 832726). Hairs were collected by cutting a few inches inward from the distal end, therefore excluding the roots. The length of hair on the head before cutting was roughly 10 cm. Hair shafts were further cut to a length of 20 mm before continuing with protein extraction. The African hair samples weighed almost half of the weight of the European hair samples due to differences in hair shaft width and shape (data not shown). Hair shafts were biochemically processed using an optimized processing protocol as part of an earlier study. After initial preparation and use of the samples to generate the data-dependent acquisition datasets used in the cited study, the remaining supernatants were stored at -20°C. Prior to mass spectrometric analysis, the samples were again centrifuged to minimize insoluble particulates in the supernatant.