High resolution SNP panels for forensic identification of ancestry, family, and phenotype.

As submitted by the proposer: SNPs can provide information that the conventional STR polymorphisms (e.g. CODIS markers) cannot. In addition, genotyping of SNPs can utilize newer less expensive methodologies. Consequently, many researchers are attempting to identify and characterize SNPs for use in forensics. In 2007 we defined four types of SNP panels for use in forensics: IISNPs for individual identification, AISNPs for ancestry inference, PISNPs for phenotype (visible trait) inference, and LISNPs for lineage/family/clan inference. The different panels require SNPs with different characteristics. We have worked to identify more efficient and robust panels of SNPs for Individual Identification, Ancestry Inference, and Lineage Inference. In addition, because forensic applications of IISNP, AISNP, and LISNP panels require reference data--all interpretations are functions of population-specfic allele frequencies--we have worked to improve the reference allele frequency data for the panels we are developing on a global representation of populations. Previously, such data existed for only a few published panels. Our work on IISNPs seems complete with a published panel of 86 statistically independent SNPs with random match probabilities, irrespective of population, of <10-15 if only the unlinked 45 IISNPs are used to <10-30 if all 86 are used. We are now proposing research focused on further improvement of AISNP and LISNP panels. Results to date using a set of 55 highly selected AISNPs provide good resolution of seven to eight geographical regions in global analyses of 73 populations thereby providing a 'first tier' of resolution of individuals into populations of origin. However, many 'first tier' SNPs are not good for distinctions among populations within one of the broad geographic regions. Other SNPs, not good globally, can show distinctions among populations within a region when populations within a geographic region are analyzed as a separate group. Such SNPs will be needed for a 'second tier' refinement of distinctions within each region. We are optimizing this hierarchical approach and expect to have our 55-AISNP set of first tier markers genotyped for over a hundred populations from around the world by the time the current project ends. These data should point to an even more powerful first tier panel as well as identify the regions needing a second tier panel. The proposed AISNP research will focus on developing and optimizing the 'second tier' region-specific panels of AISNPs. With new resources for identifying candidate SNPs (e.g., 1000 Genomes) and our unique population resources and the ability to broaden geographic coverage through collaborations, we will provide much more powerful ancestry informative panels of SNPs capable of fine geographic resolution of ancestry. Because identification of the family to which a DNA profile belongs is very important in mass disasters and missing persons situations, we have been working towards better LISNPs, a panel of markers much more powerful for identifying relatives than is possible with the CODIS STRPs. We have demonstrated the feasibility of minihaplotypes (<10kb) as LISNPs and propose to move toward microhaplotypes, smaller segments of DNA (<~200bp), and sequencing technology to minimize the ambiguity of phase that exists for the current set of minihaps. This move to develop microhaplotypes that utilize more powerful sequencing technology will increase the ability to identify members of an extended family or clan and help in correctly detecting origins of individuals of mixed ancestry. The new resources for identifying microhaplotypes (1000 Genomes and other whole genome data on a population) will provide enough candidates that we can then evaluate in the context of our broad population coverage to assure a powerful panel of LISNP loci. ca/ncf