Ethnicity and cryptic relatedness was determined using Identity-by-State (IBS) clustering and multidimensional scaling analyses within PLINK using genotypes that had been merged with data from HapMap Phase III, ASW, TSI, CEU, JPT, CHB and YRI populations [http://hapmap.ncbi.nlm.nih.gov/]. The subset of SNPs used was shared across studies, using only common SNPs that are not correlated within a 50 SNP sliding window at an r2 > 0.20, with each window overlapping by 5 SNPs. Samples were clustered using multidimensional scaling, removing outliers > 3 standard deviations from the mean component vector estimates for C1 or C2 for the combined CEU and TSI samples. Cryptically related samples were excluded after pairwise identical by descent estimates were calculated, excluding any samples sharing greater than a 0.15 proportion of alleles.
