We investigated the reliability and practicality of CAUSALdb using GWAS results for coronary artery disease (CAD). By searching for the trait name ‘Coronary Artery Disease,’ we found six CAD GWAS summary statistics from the EUR population in the current version of CAUSALdb (Supplementary Table S5). Among them, the largest study, involving meta-analysis of the UKBB and CARDIoGRAMplusC4D samples (49), showed the highest number of causal blocks (n = 165). In the original publication, the authors performed GWAS fine-mapping on 161 CAD risk loci using PAINTOR (26). Although they only considered variants having r2 >0.1 with the leading variant and GWAS P-value of <0.01 as independent loci, we found highly consistent PP of potential causal variants between the original results and CAUSALdb (Pearson correlation coefficient: PAINTOR = 0.922, CAVIARBF = 0.920, FINEMAP = 0.921; Supplementary Figure S9A). For example, CAUSALdb recapitulated 100% (8/8) variants with PP equal to 1 and 92% (69/75) variants with PP >0.5. These variants with high PP were very easy to distinguish from others and were more likely to be causal. One of the eight variants with
