We found that CUD severity and the schizophrenia PRS were not significantly associated in the European ancestry nor the African ancestry samples of COGA (ANOVA F-statistic = 1.73 [3, 2919.7], P = .158 in the European ancestry sample; F-statistic = 0.86 [3, 1596.7], P = .462 in the African ancestry sample). Associations between the schizophrenia PRS and cannabis-related experiences were attenuated but still significant when controlling for age at first use, except for paranoia, which no longer passed our statistical significance threshold (P = 8.5e−4; supplementary table 2). Similarly, previous associations were still significant when controlling for AUD diagnosis except for the association with cannabis-related paranoia (P = .001; supplementary table 3). When we controlled for a PRS for CUD in the models, none of the associations with the schizophrenia PRS were attenuated (supplementary table 4; note, this sensitivity analysis was only performed in the European ancestry subset of COGA). None of the interactions between the schizophrenia PRS and duration of daily cannabis use passed our significance threshold of α = 9.8e−4 (P > .42). The same was true for
