GWAS was performed using PLINK v1.9 (ref. 29) testing for associations between SNPs and alcohol consumption in unrelated individuals with location of UKB assessment centre, genotyping batch and 15 principal components included as covariates. In order to distinguish independent GWAS signals, SNPs with an association P-value ⩽1 × 10−4 were subjected to linkage disequilibrium (LD)-based clumping that was performed in PLINK,29 using an LD r2 cutoff of 0.2 and a 500 kb sliding window. Conditional SNP analyses were performed using PLINK for the 8 SNPs on chromosome 4q found to be associated with alcohol consumption. Each SNP was retested for association using the same method as described for the GWAS but using one of the seven other SNPs as a covariate. Conditional SNP analyses were also performed in GCTA using a stepwise model selection procedure to confirm which SNPs were independently associated.30, 31
