Genotype calling, QC and imputation proceeded separately by gene chip set. Normalized bead intensity data obtained for each sample were called using Illumina Genome Studio with cluster position files provided by Illumina, and fluorescence intensities were converted into SNP genotypes separately for each gene chip. Initial QC was performed using PLINK17 to remove markers with: zero alternate alleles, genotyping call rate ≤ 0.98, Hardy-Weinberg P value < 5 × 10−5, and individuals with genotyping call rate < 0.90. Samples were then imputed to HRC (r1.1 2016)18, as follows: if necessary, marker positions were lifted-over to GRCh37 and aligned to the HRC loci using HRC-1000G-check-bim-v4.2 (https://www.well.ox.ac.uk/~wrayner/tools/), which checks the strand, alleles, position, reference/alternate allele assignments and frequencies of the markers, removing A/T & G/C single nucleotide polymorphisms (SNPs) with minor allele frequency (MAF) > 0.4, SNPs with differing alleles, SNPs with > 0.2 allele frequency difference between the genotyped samples and the HRC samples, and SNPs not in reference panel. Imputation was performed via the Michigan Imputation Server19 using Eagle v2.320 as the phasing algorithm. Since all imputation was performed using the GRCh37, the genotype data was subsequently lifted over to GRCh38 for downstream integrating with functional genomics assays.
