Methods to detect population stratification evaluate data from a set of randomly selected loci (AIMS) not associated with the trait of interest for patterns of association indicating the likelihood of strata within a sample. We used the Structured Association (SA) method of Pritchard and colleagues (Pritchard, Stephens, & Donnelly, 2000) to adjust for stratification. This procedure uses a Bayesian clustering approach to identify populations within a sample on the basis of genotypic information at a set of markers in linkage equilibrium that are unlinked to the phenotype of interest. The procedure assigns individuals probabilistically to a population based on their genotypes. This assignment permits modeling of the candidate gene effect within populations without the need to assume that the effect is the same in all populations by testing for association conditional on the inferred ancestries of the individuals (Pritchard, Stephens, Rosenberg, & Donnelly, 2000). Our analysis revealed three distinct clusters within our population which we identified as African-American, Caucasian, and Other. Only three individuals were clustered in a group different than that of their self-reported race. Subsequent analyses revealed no
