We conducted MR analyses to elucidate the potential causal impact of SI and CPD on BMI, schizophrenia and EA using the MR-Egger, median weighted and inverse variance weighted methods. We found a causal association between SI and EA using both the median weighted and inverse variance weighted methods (P < 0.0001; Supp. Table 7) but not with MR-Egger (P = 0.2). There was an association of SI with BMI using MR-Egger only (P = 0.01; Supp. Table 7), but there was evidence of horizontal pleiotropy (P = 0.001) and no support from the other methods. Similarly, increased CPD was only associated with reduced BMI using the weighted median approach (P = 0.009) and not the other methods (P > 0.017). We also tested if schizophrenia, EA or BMI causally influence CPD or SI using SNVs associated with schizophrenia, EA and BMI, respectively, as instrumental variables. No evidence of such reverse causation was found (Supp. Table 7). These results were consistent with previous analyses [50]. There was no evidence of a causal effect of SI on schizophrenia, or CPD on educational attainment (Supp. Table 7).
