In order to compare results at the SNP level, we imputed ~2.5M common SNPs included in HapMap, using the HapMap phase II CEU data as the reference sample. Most studies used NCBI build 36 (UCSC hg18), although in the NTR/NESDA study build 35 (UCSC hg17) was used. Imputation was carried out using IMPUTE for the NTR/NESDA, SAGE and EGP samples (consisting of unrelated individuals).61 For all other samples, genotype data were imputed using MACH software. For those studies that contained related individuals, a maximum likelihood approach was used that takes advantage of the relatedness among individuals.62 Throughout this paper, the location of SNPs reported is taken from the build 36 (release 22) HapMap data.
