Description of original award (Fiscal Year 2015, $397,337)
As submitted by the proposer:
A common component of evidence at crime scenes, hair is a biologically and anatomically complex tissue that can provide a significant amount of information. The anatomy of hair shows distinctive characteristics at the level of an individual and as a larger genetic group. Unfortunately the complexity of anatomical features and the degree of morphological variation is too great to be useful as a means of individualization. The level of weighting for each physical characteristic is intrinsically subjective and results in difficulty in presenting associations in a statistically sound and scientific manner. Even distinctive traits such as racial background are difficult to present systematically.
Rather than subjectively inferring the biogeographic background of the donor, this proposal focuses on genetically variant peptides in the hair proteome. These peptides contain single amino acid polymorphisms, the result of non-synonymous single nucleotide polymorphisms (nsSNPs). Their presence in a proteomic dataset is indicative of the presence of the corresponding nsSNPs in the donors genome. The distribution of these nsSNPs often varies widely among different populations. The overall profile probability of occurrence in different populations, such as the African and European sample populations in the 1000 Genomes Project, will differ as a function of the genetic background of the individual.
We seek to establish a training set of 120 hair proteomes from different genetic backgrounds, comprising known ancestral backgrounds (from Ancestry Informative Markers) and likelihood values resulting from profile probabilities in different populations. The cohort will comprise 50 European Americans, 25 African Americans and 25 Biracial African Europeans. We will then establish a calibration curve and test it by predicting the ancestry of 20 individuals with a range of genetic admixture.
Predicting phenotype from genotype has a firm precedent in forensics, and a range of biostatistics strategies for this purpose are available to the investigator. The likelihood values we have obtained to this point range over 6 orders of magnitude, indicating that establishing a scientifically sound, statistically solid, laboratory based method to predict the biogeographic background of the hair shaft donor is within our capability. By refining the method to use a single hair, we anticipate that this method will soon be available to provide valuable information for investigators.
This project contains a research and/or development component, as defined in applicable law.