To reduce this bias, we also focused on SMR results from MTAG using stricter threshold (GWAS AUD 5 × 10−5; e/mQTL P 5 × 10−8; SMR P < 20%; Heidi P > 0.05). The results from this threshold were similar to our original analyses that focused on overlapping results after integration analysis (Supplementary Data 7 and Supplementary Data 9). In fact, many SNPs in the combined analysis were filtered out due to inflated summary statistics of the heterogeneity (Heidi P values) test (Supplementary Data 9).
