ABSTRACT
Despite the ever-widening choice of SNP genotyping systems available in the vast eld of genetic analysis, the strategy adopted by the majority of association studies applied to medical genetics research has centered on a consensus approach that has relevance for many aspects of forensic SNP typing. First, the ancestry of study subjects can be assessed, generally using panels comprising sets of small multiplexes of SNPs highly informative for ancestry (AIM-SNPs) to ensure that case and control groups are not stratied for ancestry-such stratication potentially creates false associations between any population-differentiated loci and the disease or trait investigated (Campbell et al. 2005; Marchini et al. 2004). Rebalancing the ancestry compositions of case and control groups by rearranging or excluding individuals that are incorrectly assigned to a population can save time
and money by focusing the expensive genome-wide high-density SNP typing on the most useful study donors (Sladek et al. 2007). e important point for forensic SNP analysis is the potential application of such AIM-SNPs for ancestry inference. If the AIM-SNPs are well dierentiated across populations, then relatively small-scale typing approaches can be used and forensic analysis can benet from the publication of marker lists primarily designed for association study purposes. Following ancestry checks, whole-genome scans provide genome-wide screens for association (Phase I studies) and this provides another key source of data for forensic analysis. In the last 10 years, a comprehensive range of association studies has pinpointed genomic regions strongly associated with particular traits (Donnelly 2008; Hirschhorn and Gajdos 2011). is can potentially accelerate the building of SNP sets able to predict common physical characteristics, with coding or promoter SNPs strongly associated with their expression. Identication of the key SNPs underlying the studied traits represents the last stage (Phase II) of association studies where very small numbers of SNPs are typed in much larger cohorts of study subjects to conrm/reject the associations or nd new low-frequency SNPs. Here the choice of small-scale typing platforms can feed new technologies into forensic analysis, although many high-throughput systems used for Phase II studies still depend on prohibitively large amounts of input DNA.
