found weak evidence for a causal effect of cannabis use to schizophrenia and much stronger evidence for a causal effect in the other direction. Our findings may indicate that individuals at risk for developing schizophrenia experience prodromal symptoms or negative affect that make them more likely to start using cannabis to cope or self-medicate48. The lack of strong evidence of a causal influence of cannabis use on schizophrenia may be due to the lower power of the instrumental variables. The instrumental variable based on schizophrenia SNPs explained 3.38% of variance in liability to schizophrenia. For cannabis use, the genetic instruments explained 1.12% and 0.15% of the variance in cannabis use for SNPs included with p<1e-05 and p<5e-08, respectively.
