Controlling for GENSUB revealed that the majority of nominally significant substance-disorder relationships were driven by associations between PRS and general substance involvement liability, though a few substance-disorder pairings appear to be specific: ADHD PRS with all levels of nicotine involvement and non-problem cannabis use; BIP PRS with non-problem cannabis use; MDD PRS with alcohol dependence relative to use with mild problems, as well as severe cocaine dependence relative to never-use and non-problem use, and severe cannabis dependence relative to never-use (though, contrary to expectations, this residual association was negative); and SCZ PRS with non-problem cannabis use as well as severe cocaine dependence relative to non-use and use with 0 and 3–5 dependence symptoms.
