LDSC analyses demonstrated moderate to strong genetic correlations between OA and a variety of substance use, psychiatric, and cognitive phenotypes in expected directions (e.g., positive correlation with cannabis use disorder, inverse correlation with age of smoking initiation). Moving from the general genomic signal to the specific OA-associated genes, we observed important differences in OPRM1, PPP6C, and FURIN associations with brain- and SUD-related phenotypes. Although OPRM1 has been broadly studied, only GWAS of OUD58 and methadone dose59 have identified GWS associations with OPRM1 variants (Supplementary Fig. 18; Supplementary Table 19a)60,61. The variant associated with methadone dose, rs73568641, was not associated with OA in this gSEM GWAS (p = 0.328). In contrast with OPRM1, variants in PPP6C have been associated with numerous brain- and SUD-related phenotypes, notably opioid medication use, alcohol consumption, numerous smoking phenotypes, and depression among others (Supplementary Fig. 19; Supplementary Table 19b). The specific PPP6C variants nominally associated with OA in this gSEM GWAS have been associated with neuroticism, depressive symptoms and a number of smoking phenotypes at genome-wide significance (Supplementary Table 19b). Variants in FURIN have been associated
