GWAS of Bipolar 1 Disorder in a Multi-ethnic Cohort of 72,823 Identifies Four Novel Loci. C. Schaefer1, L. Shen1, K. Kearney1, M. McCormick2, S. P. Hamilton3, L. A. McInnes3, V. Reus4, J. Wall5, P.-Y. Kwok5, M. Kvale5, T. J. Hoffmann5, E. Jorgenson1, N. Risch1,5 1) RPGEH, Division of Research, Kaiser Permanente, Oakland, CA; 2) Dept. of Psychiatry, Kaiser Permanente Santa Clara, Santa Clara, CA; 3) Dept. of Psychiatry, Kaiser Permanente San Francisco, San Francisco, CA; 4) Dept. of Psychiatry, UCSF School of Medicine, San Francisco, CA,; 5) Institute for Human Genetics, UCSF, San Francisco, CA.

   Genome-wide association studies (GWAS) of bipolar disorder (BP) have identified susceptibility variants, and some have been replicated in the analysis of the Psychiatric Genetics Consortium (PGC), but the variation explained in BP is small and studies are considered under-powered. Most analyses have been performed in non-Hispanic whites. To examine potential associations with bipolar 1 disorder (BP1) in whites and other groups, we performed a GWAS and meta-analysis in a multi-ethnic sample of BP1 cases and screened controls, and examined consistency with results from the PGC. Cases and controls are participants in the Kaiser Permanente Research Program on Genes, Environment, and Health (RPGEH). Cases were initially identified with multiple treatment episodes and/or hospitalization with BP1 from electronic medical records (EMR) with diagnosis confirmed by telephone interview. Individuals were excluded with diagnoses of schizophrenia spectrum disorder (SSD), schizoaffective disorder, or nonaffective psychoses. The resulting sample was 68% female; average age at study entry was 47 years, and 25% were minority or mixed race - ethnicity. Potential controls were drawn from the RPGEH Genetic Epidemiology Research on Adult Health and Aging general cohort. Controls were screened using EMR data to exclude those with one or more diagnosis of Axis 1 or other selected neuropsychiatric disorders. Cases and controls were genotyped using the Affymetrix Axiom system; four custom ancestry-specific arrays with 675,000 to 891,000 SNP markers were used to genotype the non-Hispanic white, Asian, African-American, and Latino participants separately. Following quality control and filtering, case-control analyses of non-Hispanic whites (3726 cases; 54620 controls) identified multiple variants in four regions with genome-wide significance (p < 5.0E-08): FAM160A1 on chr 4; SLC44A4 and CFB on chr 6; ENOX1 on chr 13; and RBFOX1 on chr 16. Analyses controlled for age, gender, and genetic ancestry. In analysis of the other race-ethnicity groups, SNPs on chr 13 were consistently associated with BP1 in the other groups, despite relatively small sample sizes. Our results generally do not replicate the significant associations found in the PGC meta-analysis, although two of the SNPs we identified on Chr 13 were associated with BPD in the PGC study.

You may contact the first author (during and after the meeting) at