To assess potential gender differences in the association signal, we tested for SNP × Sex interaction using logistic regression, coding each SNP in an additive framework, including gender and a SNP × Sex interaction term. To test for differences in evidence for association between sets of cases defined by FH status, we stratified the sample by the presence (FH+) or absence (FH−) of first degree relatives with schizophrenia or unspecified psychosis as described above (1st degree broad definition). Genotypes at rs1344706 were missing for 23 individuals with FH data, and We tested for allelic (1df) association by assessing Χ2 between FH+ and FH− cases. Both analyses were performed using SAS42.
