our results are minimally affected by population stratification (Supplementary Note 2 and Supplementary Figs. 5–7).Table 1 Summary of results from within-ancestry and trans-ancestry GWAS meta-analysesCohort ancestry or ethnic groupNumber of studiesMax n (mean n)Number of GWS COJO SNPs (PGWAS < 5 × 10−8)Number of GWS loci (35 kb)Cumulative length of non-overlapping GWS loci in Mb (% of genome)European (EUR)1734,080,687 (3,612,229)9,863 (8,382)6,386552.5 (18.4%)East Asian (EAS)56472,730 (320,570)918 (807)82160.5 (2.0%)Hispanic (HIS)11455,180 (431,645)1,511 (1,195)1,373101.0 (3.3%)African (AFR)29293,593 (222,981)453 (404)41230.4 (1.0%)South Asian (SAS)1277,890 (59,420)69 (65)664.7 (0.2%)Trans-ancestry meta-analysis (METAFE)2815,314,291* (4,611,160)12,111 (9,920)7,209647.5 (21.6%)n denotes the sample size for each SNP. GWS: genome-wide significant (P < 5 × 10−8). COJO SNPs: near-independent GWS SNPs identified using an approximate COJO analysis implemented in the GCTA software. PGWAS: P value from a marginal association test. GWS loci were defined as genomic regions centred around each GWS SNP and including all SNPs within 35 kb on each side of the lead GWS SNP. Overlapping GWS loci were merged so that the number and cumulative length of GWS loci are calculated on non-overlapping GWS loci. The percentage of the genome covered was calculated by dividing the cumulative of GWS loci by 3,039 Mb (the approximated length of the human genome).*The number of individuals
