A potential confounder of these analyses is smoking behavior itself: because individuals with schizophrenia are more likely to smoke than controls (Hartz et al., 2014), some of the GWAS associations for schizophrenia may, in fact, be associations with smoking behavior. Because (1) there is a significant proportion of individuals with schizophrenia who do not smoke, and (2) we are looking at composite GWAS results rather than individual SNPs, it is unlikely that this confounder biases the findings.
