The meta-analysis of GS:SFHS and UKB was conducted using the classical inverse-variance approach, which weights effect sizes by sampling distribution, implemented in the METAL package39. SNPs with a meta-analysis P-value of P ≤ 1E-05 were subjected to clump-based linkage disequilibrium pruning using PLINK25 using an LD r2 cut off of 0.1 and a 500 kb sliding window to create SNP sets of approximately independent “lead SNPs”. All SNPs which surpassed genome-wide significance were entered into the NHGRI-EBI catalog of published GWAS40,41 (www.ebi.ac.uk/gwas/) to observe whether these SNPs had been previously observed in association analysis.
