68-1

9/98

for psychiatric disorders remained known. Partici-pants in our study were recruited through Bipolar Disorder Research Network Japan (BDRNJ, http://bipolar.umin.jp/), a network of patients, families, and researchers across Japan. Thanks to BDRNJ, we could report de novo mutation analysis with 354 families with bipolar disorder56). This number is the world’s largest number for de novo mutation anal-ysis in bipolar disorder to date. Research networks consisting of the patients and researchers such as BDRNJ are influential players in current genomic research. Among the extremely rare de novo mutations in bipolar disorder, protein-truncating (loss-of-func-tion) de novo mutation in a high probability of loss-of-function intolerance (pLI) genes is signifi-cantly more observed in bipolar disorder than in control. High pLI genes are the genes in which loss-of-function mutation is less observed in the general population than the theoretical estimation, meaning its susceptibility to natural selection by loss-of-function. This result is naturally expected for severe psychiatric disorders and is a common observation with schizophrenia and ASD24, 40, 61). Among the extremely rare de novo mutations, deleterious mutations are enriched in the genes related to synapse and calcium ions. This result is consistent with the results of common variants and other biological studies on bipolar disorder42, 62, 63), supporting the dysregulation of synapse and calcium signaling as the promising candidate mech-anisms of bipolar disorder.Among the genes hit by deleterious mutations found in this study, KMT2C is known to cause a severe neurodevelopmental disorder (Kleefstra syndrome64)) much more severe than bipolar disorder. We investigated the detail of the mutation in KMT2C to explain the apparent phenotypic difference between bipolar disorder and Kleefstra syndrome. Surprisingly, the mutation is, in fact, a mosaic mutation (postzygotic de novo mutation) in the proband’s body. The mutation exists only in some cells and should have occurred in the process of early development. Encouraged by this finding, we investigated postzygotic de novo mutations in patients with bipolar disorder. Among the postzy-gotic de novo mutations in bipolar disorder, delete-rious mutations are enriched in the genes known to cause neurodevelopmental disorders (e.g., KMT2C). In addition, we found two deleterious mutations in SRCAP in two independent patients with bipolar disorder. SRCAP is known to cause a severe neuro-developmental disorder, Floating-Harbor syndrome65). This result leads to an interesting hypothesis: mosaic mutations in neurodevelopmental disorder genes cause milder phenotypes, including bipolar disorder. If this hypothesis holds true, bipolar disorder shares pathological mechanisms with severe neurodevelopmental disorders. Encouraged by rare variant studies on developmental disorders to understand the pathological mechanisms of psychiatric disorders66), we are now proceeding with mosaic mutation analysis to confirm this hypothesis67).We conducted a comprehensive study of two types of mutations, germline and postzygotic (mosaic) de novo mutations, and elucidated a part of the genetic architecture of bipolar disorder56). However, analysis with more families and patients with bipolar disorder is necessary to obtain more reliable and more profound knowledge. Future genomic research will be more empowered by research networks consisting of the patients and researchers such as BDRNJ. Through such collabo-ration, we will understand the pathological mecha-nisms and develop a new therapeutic/preventive strategy for bipolar disorder.Future direction: genotype first approach In this Review, I have overviewed the current understanding of major psychiatric disorders by genomic investigation. Genomic analysis is one of the most powerful approaches to elucidate the biological mechanisms of psychiatric disorders. However, accumulated data from genomic studies so far proposes a critical question to the current diagnostic boundary among psychiatric disorders. Some genes are simultaneously associated with different disorders through common and rare vari-ants47-49). This is particularly notable for common variants associated with bipolar disorder and schizophrenia. In addition to the shared risk of common variants, AKAP11, the first gene reported to be associated with bipolar disorder through a gigantic meta-analysis of exome sequencing for rare variants, is also associated with schizo-phrenia68). These results suggest that major psychi-7in psychiatry