Fig. 3: Cross-ancestry effect-size comparison and genetic correlation analysis.
a, Per-allele effect-size (logarithm of ORs) comparison using east Asian-specific and European-specific summary statistics for PUD. Lead variants associated with PUD or any subtype in east Asian-specific, European-specific or cross-ancestry meta-analysis were selected for comparison. Two-sided P values were derived from the ancestry-specific meta-analyses. The most significant associations were shown if overlapping variants existed (interval <500 kb). Data are presented as effect-size estimates (log(OR)) ± s.e.m. Variants with nominally significant heterogeneity (Cochran’s Q test; two-sided Phet < 0.05) were denoted by the black marker edges. The gray dashed line represents the fitted linear regression line with annotation at the bottom right (P value is derived from a two-sided Student’s t-test for the slope). Pearson’s r is shown. b, Cross-ancestry genetic correlation for PUD, GU and DU estimated by Popcorn. The gray dashed line indicated the null hypothesis (ρ = 1). Asterisks indicate estimates that were significantly <1 (two sided; FDR < 5%). Data are presented as genetic correlation estimates ± s.e.m. Sample sizes are provided in Supplementary Table 12.
