68-1

8/98

6from severe depression and problematic behavior from manic moods significantly impact patients’ social life. Elucidating the pathophysiology of bipolar disorder and the development of new treat-ments are required42). Similar to the situation of schizophrenia, the prevalence of around 1% is on the threshold of common or rare diseases. Genomic investigation of bipolar disorder is also advancing under the two models: research to common vari-ants as common diseases and rare variants as rare diseases.Studies on common variants precede studies on rare variants in the genetics of bipolar disorder43-46). GWASs on bipolar disorder have detected 64 loci associated with bipolar disorder using a consor-tium-based approach with aggregated 42,000 cases and 370,000 controls44). As with the situation in schizophrenia, the effect size of each variant is rela-tively small. The odds ratio of the associated loci in bipolar is 1.15 at most, which is even smaller than schizophrenia. However, the number of the associ-ation is biologically informative, elucidating the gene sets related to synapse and ion channels. Synaptic structure and function are relevant to both bipolar disorder and schizophrenia. The phenotypes of bipolar disorder and schizophrenia Figure 2　A schematic illustration of the two classes of de novo mutations: germline and postzygotic de novo mutations. The germline de novo mutations arise in the gametogenesis of the parents. The postzygotic de novo mutations arise in the developmental process of the proband after the fertilization.are apparently different, but the correlation of asso-ciated common variants is notable between bipolar disorder and schizophrenia with a genetic correla-tion of around 0.6844, 47-49). This high correlation is probably against most clinical psychiatrists’ intu-ition and should be the focus of future psychiatric genomics50).In contrast to studies on common variants, studies on rare variants in bipolar disorder lags behind in sample size despite several pioneering studies51-55). Therefore, we are investigating rare variants, especially de novo mutations, to elucidate the genetic architecture of bipolar disorder53, 56). De novo mutations are subject to little natural selec-tion and are thought to contain disease-associated mutations with significant effects. In particular, we are investigating extremely rare de novo muta-tions not found in the general population to find potential disease-associated rare variants56). We also investigated postzygotic de novo mutations (mosaic or somatic mutations) in addition to clas-sical de novo mutations (i.e., germline de novo mutations) to explore the unknown genetic archi-tecture of bipolar disorder (Figure 2). While several studies have reported on mosaic mutations in ASD57-60), pathological roles of mosaic mutations