SNP genotype imputation was performed in the Yale-Penn and the SAGE GWAS datasets with IMPUTE2 (21) using genotyped SNPs with a minor allele frequency of >1% and the June 2011 1000 Genomes reference panel (22), which contains phased haplotypes for 1094 individuals of various ancestries: 379 of European descent (CEU, FIN (Finnish in Finland), GBR (British from England and Scotland), IBS (Iberian populations in Spain), and TSI (Toscani in Italia)), 286 of Asian descent CHB, JPT (Japanese in Tokyo, Japan), and CHS (Han Chinese South, China), 181 admixed American samples (PUR (Puerto Rican in Puerto Rico), CLM (Colombian in Medellin, Colombia), and MXL (Mexican ancestry in Los Angeles, California)), and 246 samples of African descent (ASW (African ancestry in southwest USA), LWK (Luhya in Webuye, Kenya), YRI) (22). All samples were imputed using every available sample in the reference panel, then split into AA and EA datasets based on the clustering techniques described above. We retained 18,564,419 SNPs with derived information content >.8 in at least one of the population groups. After excluding SNPs with MAF < 3% in both
