Disrupted in schizophrenia 1 (DISC1) has been associated with risk of schizophrenia, schizoaffective disorder, bipolar disorder, major depression, autism and Asperger syndrome, but apart from in the original translocation family, true causal variants have yet to be confirmed. Here we report a harmonized association study for DISC1 in European cohorts of schizophrenia and bipolar disorder. We identify regions of significant association, demonstrate allele frequency heterogeneity and provide preliminary evidence for modifying interplay between variants. Whereas no associations survived permutation analysis in the combined data set, significant corrected associations were observed for bipolar disorder at rs1538979 in the Finnish cohorts (uncorrected P=0.00020; corrected P=0.016; odds ratio=2.73+/-95% confidence interval (CI) 1.42-5.27) and at rs821577 in the London cohort (uncorrected P=0.00070; corrected P=0.040; odds ratio=1.64+/-95% CI 1.23-2.19). The rs821577 single nucleotide polymorphism (SNP) showed evidence for increased risk within the combined European cohorts (odds ratio=1.27+/-95% CI 1.07-1.51), even though significant corrected association was not detected (uncorrected P=0.0058; corrected P=0.28). After conditioning the European data set on the two risk alleles, reanalysis revealed a third significant SNP association (uncorrected P=0.00050; corrected P=0.025). This SNP showed evidence for interplay, either increasing or decreasing risk, dependent upon the presence or absence of rs1538979 or rs821577. These findings provide further support for the role of DISC1 in psychiatric illness and demonstrate the presence of locus heterogeneity, with the effect that clinically relevant genetic variants may go undetected by standard analysis of combined cohorts.