In the discovery GWAS, we meta-analyzed Yale-Penn 1 and 2 EAs and AAs separately by population (Fig. 1). In the meta-analysis of Yale-Penn EAs, three GWS SNPs representing one signal were identified (see GWAS results in Fig. 1 for EAs): rs115455482, rs74722579, and rs1938228, which are highly correlated (r2 ≥ 0.95, MAF ~ 0.06 in EAs when cases and controls were considered together; Table 1 and Supplementary Table S1). The association P-values were 1.66 × 10−10, 2.81 × 10−9, and 1.57 × 10−8 for rs115455482, rs74722579, and rs1938228, respectively; the leading SNP rs115455482 is ~230 kb upstream of transcript NM_000740 of CHRM3 (muscarinic acetylcholine receptor (AChR) M3). There were no GWS SNPs identified in AAs (Supplementary Figs. 1 and 2 for AAs). Low inflation was observed in the meta-analysis of Yale-Penn 1 and 2 for EAs (λ = 1.06; Supplementary Fig. 3) and AAs (λ = 1.02; Supplementary Fig. 2).Table 1Meta-analysis of rs115455482 and rs74722579 in European American (EA) and African American (AA) long-term cannabis users.PopCohortaHallucination control no.Hallucination case no.SNPbAlleleMAFcEffect β (SE)dPCaseControlEAYale-Penn 1104151rs115455482T0.1670.0400.13 (0.03)4.93 × 10−9rs74722579C0.1770.0480.12 (0.03)1.66 × 10−8Yale-Penn 278540rs115455482T0.0880.0310.08
