Autosomal genotype data were available for all GS:SFHS individuals in the present study (n = 18 725). Genotypes were imputed using the Haplotype Reference Consortium reference panel (HRC.r1-1)19 via the Sanger Imputation Server pipeline (https://imputation.sanger.ac.uk). Prior to imputation, individuals with missingness ≥ 3% were excluded, as were SNPs with a call rate of ≤98%, Hardy Weinberg Equilibrium (HWE) P-value ≤ 1 × 10−6, and a minor allele frequency (MAF) ≤ 1%. Phasing of genotype data was performed using the SHAPEIT2 algorithm20 utilizing the duoHMM option, which refines phasing by utilizing pedigree information. Imputation was performed using PBWT software21. Multi-allelic variants, monomorphic variants and SNPs with an imputation INFO score < 0.8 were removed22. Population outliers (more than 6SDs from the mean of the first principal component (PC)) were identified and removed from the sample23, as were one from each of 52 monozygotic twin pairs, identified by IBD (preferentially retaining cases), and 7 individuals who matched samples from the Psychiatric Genomics Consortium, identified using genotype checksums24. After imputation, individuals with missingness ≥ 2%, and genotype with a call rate of ≤98%,
