Cross-Disorder Genomewide Analysis of Schizophrenia, Bipolar Disorder, and Depression

Objective: Family and twin studies indicate substantial overlap of genetic influences on psychotic and mood disorders. Linkage and candidate gene studies have also suggested overlap across schizophrenia, bipolar disorder, and major depressive disorder. The purpose of this study was to apply genomewide association study (GWAS) analysis to address the specificity of genetic effects on these disorders. Method: The authors combined GWAS data from three large effectiveness studies of schizophrenia (CATIE, genotyped: N = 741), bipolar disorder (STEP-BD, genotyped: N = 1,575), and major depressive disorder (STAR*D, genotyped: N = 1,938) as well as from psychiatrically screened control subjects (NIMH-Genetics Repository: N = 1,204). A two-stage analytic procedure involving an omnibus test of allele frequency differences among case and control groups was applied, followed by a model selection step to identify the best-fitting model of allelic effects across disorders. Results: The strongest result was seen for a single nucleotide polymorphism near the adrenomedullin (ADM) gene (rs6484218), with the best-fitting model indicating that the effect was specific to bipolar II disorder. Findings also revealed evidence suggesting that several genes may have effects that transcend clinical diagnostic boundaries, including variants in NPAS3 that showed pleiotropic effects across schizophrenia, bipolar disorder, and major depressive disorder. Conclusions: This study provides the first genomewide significant evidence implicating variants near the ADM gene on chromosome 11p15 in psychopathology, with effects that appear to be specific to bipolar II disorder. Although genomewide significant evidence of cross-disorder effects was not detected, the results provide evidence that there are both pleiotropic and disorder-specific effects on major mental illness and illustrate an approach to dissecting the genetic basis of mood and psychotic disorders that can inform future largescale cross-disorder GWAS analyses.