Advertisement

Cross-Disorder Analysis of Genic and Regulatory Copy Number Variations in Bipolar Disorder, Schizophrenia, and Autism Spectrum Disorder

Open AccessPublished:April 21, 2022DOI:https://doi.org/10.1016/j.biopsych.2022.04.003

      Abstract

      Background

      We aimed to determine the similarities and differences in the roles of genic and regulatory copy number variations (CNVs) in bipolar disorder (BD), schizophrenia (SCZ), and autism spectrum disorder (ASD).

      Methods

      Based on high-resolution CNV data from 8708 Japanese samples, we performed to our knowledge the largest cross-disorder analysis of genic and regulatory CNVs in BD, SCZ, and ASD.

      Results

      In genic CNVs, we found an increased burden of smaller (<100 kb) exonic deletions in BD, which contrasted with the highest burden of larger (>500 kb) exonic CNVs in SCZ/ASD. Pathogenic CNVs linked to neurodevelopmental disorders were significantly associated with the risk for each disorder, but BD and SCZ/ASD differed in terms of the effect size (smaller in BD) and subtype distribution of CNVs linked to neurodevelopmental disorders. We identified 3 synaptic genes (DLG2, PCDH15, and ASTN2) as risk factors for BD. Whereas gene set analysis showed that BD-associated pathways were restricted to chromatin biology, SCZ and ASD involved more extensive and similar pathways. Nevertheless, a correlation analysis of gene set results indicated weak but significant pathway similarities between BD and SCZ or ASD (r = 0.25–0.31). In SCZ and ASD, but not BD, CNVs were significantly enriched in enhancers and promoters in brain tissue.

      Conclusions

      BD and SCZ/ASD differ in terms of CNV burden, characteristics of CNVs linked to neurodevelopmental disorders, and regulatory CNVs. On the other hand, they have shared molecular mechanisms, including chromatin biology. The BD risk genes identified here could provide insight into the pathogenesis of BD.

      Keywords

      Although bipolar disorder (BD), schizophrenia (SCZ), and autism spectrum disorder (ASD) have traditionally been considered separate disease entities, they share some common behavioral characteristics and cognitive deficits. Genetic epidemiological studies have suggested the presence of shared genetic factors among these and other psychiatric disorders (
      • Sullivan P.F.
      • Magnusson C.
      • Reichenberg A.
      • Boman M.
      • Dalman C.
      • Davidson M.
      • et al.
      Family history of schizophrenia and bipolar disorder as risk factors for autism.
      ,
      • Song J.
      • Bergen S.E.
      • Kuja-Halkola R.
      • Larsson H.
      • Landen M.
      • Lichtenstein P.
      Bipolar disorder and its relation to major psychiatric disorders: A family-based study in the Swedish population.
      ,
      • Stahlberg O.
      • Soderstrom H.
      • Rastam M.
      • Gillberg C.
      Bipolar disorder, schizophrenia, and other psychotic disorders in adults with childhood onset AD/HD and/or autism spectrum disorders.
      ). In line with this, rare copy number variations (CNVs) at multiple loci have been identified as shared risk factors for SCZ and ASD (
      • Pinto D.
      • Delaby E.
      • Merico D.
      • Barbosa M.
      • Merikangas A.
      • Klei L.
      • et al.
      Convergence of genes and cellular pathways dysregulated in autism spectrum disorders.
      ,
      • Marshall C.R.
      • Howrigan D.P.
      • Merico D.
      • Thiruvahindrapuram B.
      • Wu W.
      • Greer D.S.
      • et al.
      Contribution of copy number variants to schizophrenia from a genome-wide study of 41,321 subjects.
      ,
      • Szatkiewicz J.P.
      • O’Dushlaine C.
      • Chen G.
      • Chambert K.
      • Moran J.L.
      • Neale B.M.
      • et al.
      Copy number variation in schizophrenia in Sweden.
      ,
      • Sebat J.
      • Lakshmi B.
      • Malhotra D.
      • Troge J.
      • Lese-Martin C.
      • Walsh T.
      • et al.
      Strong association of de novo copy number mutations with autism.
      ,
      • Xu B.
      • Roos J.L.
      • Levy S.
      • van Rensburg E.J.
      • Gogos J.A.
      • Karayiorgou M.
      Strong association of de novo copy number mutations with sporadic schizophrenia.
      ,
      • Kushima I.
      • Aleksic B.
      • Nakatochi M.
      • Shimamura T.
      • Okada T.
      • Uno Y.
      • et al.
      Comparative analyses of copy-number variation in autism spectrum disorder and schizophrenia reveal etiological overlap and biological insights.
      ,
      • Nakatochi M.
      • Kushima I.
      • Ozaki N.
      Implications of germline copy-number variations in psychiatric disorders: Review of large-scale genetic studies.
      ). Gene set analyses of genes affected by CNVs have implicated common molecular mechanisms in the pathogenesis of SCZ and ASD (e.g., synapse function, fragile X mental retardation protein [FMRP] targets, chromatin regulation) (
      • Pinto D.
      • Delaby E.
      • Merico D.
      • Barbosa M.
      • Merikangas A.
      • Klei L.
      • et al.
      Convergence of genes and cellular pathways dysregulated in autism spectrum disorders.
      ,
      • Marshall C.R.
      • Howrigan D.P.
      • Merico D.
      • Thiruvahindrapuram B.
      • Wu W.
      • Greer D.S.
      • et al.
      Contribution of copy number variants to schizophrenia from a genome-wide study of 41,321 subjects.
      ,
      • Szatkiewicz J.P.
      • O’Dushlaine C.
      • Chen G.
      • Chambert K.
      • Moran J.L.
      • Neale B.M.
      • et al.
      Copy number variation in schizophrenia in Sweden.
      ,
      • Kushima I.
      • Aleksic B.
      • Nakatochi M.
      • Shimamura T.
      • Okada T.
      • Uno Y.
      • et al.
      Comparative analyses of copy-number variation in autism spectrum disorder and schizophrenia reveal etiological overlap and biological insights.
      ,
      • Kirov G.
      • Pocklington A.J.
      • Holmans P.
      • Ivanov D.
      • Ikeda M.
      • Ruderfer D.
      • et al.
      De novo CNV analysis implicates specific abnormalities of postsynaptic signalling complexes in the pathogenesis of schizophrenia.
      ). On the other hand, BD-associated CNVs and genes are currently limited, and molecular mechanisms resulting from CNVs to BD are less clear (
      • Green E.K.
      • Rees E.
      • Walters J.T.
      • Smith K.G.
      • Forty L.
      • Grozeva D.
      • et al.
      Copy number variation in bipolar disorder.
      ,
      • Charney A.W.
      • Stahl E.A.
      • Green E.K.
      • Chen C.Y.
      • Moran J.L.
      • Chambert K.
      • et al.
      Contribution of rare copy number variants to bipolar disorder risk is limited to schizoaffective cases.
      ,
      • Malhotra D.
      • McCarthy S.
      • Michaelson J.J.
      • Vacic V.
      • Burdick K.E.
      • Yoon S.
      • et al.
      High frequencies of de novo CNVs in bipolar disorder and schizophrenia.
      ,
      • Priebe L.
      • Degenhardt F.A.
      • Herms S.
      • Haenisch B.
      • Mattheisen M.
      • Nieratschker V.
      • et al.
      Genome-wide survey implicates the influence of copy number variants (CNVs) in the development of early-onset bipolar disorder.
      ,
      • Gordovez F.J.A.
      • McMahon F.J.
      The genetics of bipolar disorder.
      ,
      • Grozeva D.
      • Kirov G.
      • Ivanov D.
      • Jones I.R.
      • Jones L.
      • Green E.K.
      • et al.
      Rare copy number variants: A point of rarity in genetic risk for bipolar disorder and schizophrenia.
      ).
      There are several limitations in published CNV studies. First is the use of single nucleotide polymorphism arrays that cannot reliably detect small CNVs (<50 kb), which outnumber CNVs of a larger size. The detection of such small CNVs may reveal BD-associated CNVs and genes. Second, although there is growing evidence of a cross-disorder effect of CNVs (
      • Martin C.L.
      • Wain K.E.
      • Oetjens M.T.
      • Tolwinski K.
      • Palen E.
      • Hare-Harris A.
      • et al.
      Identification of neuropsychiatric copy number variants in a health care system population.
      ,
      • Yang Z.
      • Wu H.
      • Lee P.H.
      • Tsetsos F.
      • Davis L.K.
      • Yu D.
      • et al.
      Investigating shared genetic basis across Tourette syndrome and comorbid neurodevelopmental disorders along the impulsivity-compulsivity spectrum.
      ,
      • Zarrei M.
      • Burton C.L.
      • Engchuan W.
      • Young E.J.
      • Higginbotham E.J.
      • MacDonald J.R.
      • et al.
      A large data resource of genomic copy number variation across neurodevelopmental disorders.
      ,
      • McGrath L.M.
      • Yu D.
      • Marshall C.
      • Davis L.K.
      • Thiruvahindrapuram B.
      • Li B.
      • et al.
      Copy number variation in obsessive-compulsive disorder and Tourette syndrome: A cross-disorder study.
      ), there have been few studies of cross-disorder CNV analysis that simultaneously include all three disorders (BD, SCZ, and ASD). These analyses could also reveal similarities and differences in the roles of CNVs among the three disorders. Third, CNVs affecting noncoding regulatory elements have not been fully explored despite emerging evidence for their involvement in human diseases (
      • Zhang F.
      • Lupski J.R.
      Non-coding genetic variants in human disease.
      ,
      • Murakawa Y.
      • Yoshihara M.
      • Kawaji H.
      • Nishikawa M.
      • Zayed H.
      • Suzuki H.
      • et al.
      Enhanced identification of transcriptional enhancers provides mechanistic insights into diseases.
      ,
      • Spielmann M.
      • Klopocki E.
      CNVs of noncoding cis-regulatory elements in human disease.
      ). Large consortia such as Encyclopedia of DNA Elements (ENCODE) and Roadmap Epigenomics Projects have identified enhancer or promoter regions in brain tissue (
      ENCODE Project Consortium
      An integrated encyclopedia of DNA elements in the human genome.
      ,
      • Kundaje A.
      • Meuleman W.
      • Ernst J.
      • Bilenky M.
      • Yen A.
      • et al.
      Roadmap Epigenomics Consortium
      Integrative analysis of 111 reference human epigenomes.
      ). Therefore, it is expected that the role of regulatory CNVs in psychiatric disorders may be clarified.
      In the present study, we conducted the largest known cross-disorder analysis of genic and regulatory CNVs in BD, SCZ, and ASD based on high-resolution CNV data from 8708 individuals in a Japanese population. We found an increased burden of smaller (<100 kb) exonic deletions in BD, which was in contrast to the highest burden of larger (>500 kb) exonic CNVs in SCZ and ASD. Pathogenic CNVs linked to neurodevelopmental disorders (NDDs) were associated with the risk of each disorder, but BD and SCZ/ASD differed in terms of the characteristics of NDD-CNVs. Whereas gene set analysis showed that BD-associated pathways were restricted to chromatin biology, SCZ and ASD involved more extensive and similar pathways. Nevertheless, a correlation analysis of gene set results showed weak but significant similarities between BD and SCZ/ASD. Finally, in SCZ and ASD, CNVs were significantly enriched in enhancers and promoters in brain tissue.

      Methods and Materials

      Participants

      We studied 8903 Japanese individuals, including 1843 BD cases (42.2% bipolar I disorder, 53.9% bipolar II disorder, and 4.0% unknown subtype), 1236 ASD cases, 3111 SCZ cases, and 2713 psychiatrically normal controls (Table S1 in the Supplement). Whereas all BD cases have not been previously analyzed in CNV studies, some of the samples (2519 SCZ, 1132 ASD cases, 2110 controls) were included in our previous study (
      • Kushima I.
      • Aleksic B.
      • Nakatochi M.
      • Shimamura T.
      • Okada T.
      • Uno Y.
      • et al.
      Comparative analyses of copy-number variation in autism spectrum disorder and schizophrenia reveal etiological overlap and biological insights.
      ). Disorders in cases were diagnosed according to DSM-5 criteria for BD, SCZ, and ASD. Controls were selected from the general population and had no history of mental disorders based on responses to questionnaires or self-reporting. More of the participants’ characteristics are provided in Supplemental Methods in the Supplement.
      This study was approved by the ethics committee of Nagoya University and each participating institute. Written informed consent was obtained from all participants.

      Array Comparative Genomic Hybridization

      We performed CNV analysis using two types of array comparative genomic hybridization (aCGH): NimbleGen 720K Whole-Genome Tiling array (Roche NimbleGen) and Agilent SurePrint G3 Human CGH 400K (Agilent Technologies). CNV calls were made with Nexus Copy Number 9.0 (BioDiscovery) using the Fast Adaptive States Segmentation Technique 2 algorithm. The following log2 ratio thresholds were set to detect CNVs in the NimbleGen and Agilent arrays: 10–500 kb: −0.6 (deletion) and 0.4 (duplication), >500 kb: −0.4 (deletion) and 0.3 (duplication). The significance threshold to adjust the sensitivity of the segmentation algorithm was set at 1 × 10−6, and at least 3 contiguous probes were required for CNV calls. A noise-reduction algorithm for aCGH data was used for the systematic correction of artifacts caused by GC content and fragment length (
      • Lepretre F.
      • Villenet C.
      • Quief S.
      • Nibourel O.
      • Jacquemin C.
      • Troussard X.
      • et al.
      Waved aCGH: To smooth or not to smooth.
      ).
      In terms of quality control (QC), scores were calculated for each sample based on the statistical variance of the probe-to-probe log ratios. Lower QC scores indicated better-quality results. We excluded samples with QC scores >0.2, gender mismatch, and excessive autosomal CNV calls (subject QC). Next, we excluded CNV calls <10 kb; those with low probe density (<1 probe/30 kb), >70% overlap with segmental duplications, >10% overlap with CpG islands, and call p value >1 × 10−10; and those on the Y chromosome. Finally, we filtered out common CNVs (≥1% of our total samples). Large CNVs can be split by CNV-calling algorithms. To overcome this issue, adjacent CNV calls were merged using a custom script. We merged the adjacent CNVs of the same type (i.e., deletion or duplication) if they occurred in a single individual and the gap was <50% of the entire length of the newly merged CNV. We performed all statistical analyses based on rare (<1%) CNVs. All genomic locations are given in hg18 coordinates. Gene annotation was based on GENCODE Release 35. We evaluated the accuracy of CNVs identified by aCGH using a quantitative real-time polymerase chain reaction (TaqMan copy number assays) (Applied Biosystems), as previously described (
      • Kushima I.
      • Aleksic B.
      • Nakatochi M.
      • Shimamura T.
      • Shiino T.
      • Yoshimi A.
      • et al.
      High-resolution copy number variation analysis of schizophrenia in Japan.
      ).
      As we used two types of aCGH, it is important to control for batch effects in statistical analyses. To this end, we included array type as a covariate in all analyses for SCZ and ASD. The aCGH for BD cases was performed using Agilent arrays only. Therefore, statistical analyses for BD cases versus controls were performed based on the CNV data from Agilent arrays (1818 BD cases and 1847 controls).

      Genome-wide Burden Analysis

      We performed burden analyses across a range of CNV sizes (<100 kb, 100–500 kb, >500 kb) and CNV types (deletion, duplication, deletion+duplication). The burden of CNVs was measured as the number of rare exonic CNVs. Exonic CNVs were defined as overlapping with any exon of a gene. Statistical tests were performed using a logistic regression model to predict case-control status by the number of rare exonic CNVs along with array type and sex as a covariate. One-sided empirical p values were calculated based on 100,000 permutations, swapping case-control status. The p values were adjusted for multiple testing using Bonferroni correction.

      CNVs Linked to Neurodevelopmental Disorders

      We examined whether NDD-CNVs are significantly associated with risk for BD, SCZ, and ASD. We preselected 307 NDD-linked loci (265 risk genes and 42 CNV loci) (Tables S2a and S2b in the Supplement). The NDD-linked genes were selected based on the NDD-gene databases (e.g., Developmental Brain Disorder Gene Database, denovo-db, Gene4Denovo, and SFARI database) and findings in previous literature (
      • Gonzalez-Mantilla A.J.
      • Moreno-De-Luca A.
      • Ledbetter D.H.
      • Martin C.L.
      A Cross-disorder method to identify novel candidate genes for developmental brain disorders.
      ,
      • Zhao G.H.
      • Li K.K.
      • Li B.
      • Wang Z.
      • Fang Z.H.
      • Wang X.M.
      • et al.
      Gene4Denovo: An integrated database and analytic platform for de novo mutations in humans.
      ,
      • Turner T.N.
      • Yi Q.
      • Krumm N.
      • Huddleston J.
      • Hoekzema K.
      • F Stessman H.A.
      • et al.
      denovo-db: A compendium of human de novo variants.
      ,
      • Abrahams B.S.
      • Arking D.E.
      • Campbell D.B.
      • Mefford H.C.
      • Morrow E.M.
      • Weiss L.A.
      • et al.
      SFARI Gene 2.0: A community-driven knowledgebase for the autism spectrum disorders (ASDs).
      ). Their associations with NDDs are supported by strong genetic evidence (e.g., identification of de novo loss-of-function variants in multiple patients and significant association in a large-scale case-control study). The NDD-linked CNV loci were selected from our previous study (
      • Kushima I.
      • Aleksic B.
      • Nakatochi M.
      • Shimamura T.
      • Okada T.
      • Uno Y.
      • et al.
      Comparative analyses of copy-number variation in autism spectrum disorder and schizophrenia reveal etiological overlap and biological insights.
      ). Then, we identified pathogenic or likely pathogenic CNVs in these loci according to the American College of Medical Genetics guidelines (
      • Riggs E.R.
      • Andersen E.F.
      • Cherry A.M.
      • Kantarci S.
      • Kearney H.
      • Patel A.
      • et al.
      Technical standards for the interpretation and reporting of constitutional copy-number variants: A joint consensus recommendation of the American College of Medical Genetics and Genomics (ACMG) and the Clinical Genome Resource (ClinGen).
      ,
      • Brandt T.
      • Sack L.M.
      • Arjona D.
      • Tan D.
      • Mei H.
      • Cui H.
      • et al.
      Adapting ACMG/AMP sequence variant classification guidelines for single-gene copy number variants.
      ). Further details are provided in Supplemental Methods in the Supplement.
      Next, we performed association analyses in 3 ways. First, we examined the associations of all NDD-CNVs combined. Second, we tested the associations of each subtype of NDD-CNVs with at least 5 observations. Third, we tested the associations of individual NDD-CNVs with at least 5 observations.
      Statistical analyses were conducted using Firth’s bias-reduced logistic regression model, in which case versus control status was regressed on NDD-CNVs along with array type and sex as a covariate. We calculated one-sided empirical p values based on 100,000 permutations, swapping case-control status. The empirical significance values obtained via permutation are robust to data with sparse cell counts (
      • Raychaudhuri S.
      • Korn J.M.
      • McCarroll S.A.
      • Altshuler D.
      • Sklar P.
      • et al.
      International Schizophrenia Consortium
      Accurately assessing the risk of schizophrenia conferred by rare copy-number variation affecting genes with brain function.
      ). The p values were adjusted for multiple testing using Bonferroni correction.

      Gene Set Analysis

      To identify biological pathways underlying the pathogenesis of each disorder, we tested for the enrichment of rare exonic CNVs in gene sets relative to all rare exonic CNVs. Specifically, we used a logistic regression model, in which case versus control status was regressed on the number of genes within a given gene set that were intersected by rare exonic CNVs, with adjustment for covariates, including array type, sex, total length of rare CNVs, and number of rare CNVs. This method is robust against not only batch effects, but also case-control differences in total length of rare CNVs, number of rare CNVs, and systematic differences in gene size (
      • Raychaudhuri S.
      • Korn J.M.
      • McCarroll S.A.
      • Altshuler D.
      • Sklar P.
      • et al.
      International Schizophrenia Consortium
      Accurately assessing the risk of schizophrenia conferred by rare copy-number variation affecting genes with brain function.
      ). The enrichment in cases was reported as one-sided empirical p values using 100,000 permutations, swapping case-control status. Multiple-testing correction was performed separately for each gene set group and CNV type using the Benjamini-Hochberg false discovery rate (
      • Benjamini Y.
      • Hochberg Y.
      Controlling the false discovery rate—a practical and powerful approach to multiple testing.
      ). Gene sets were considered significant if the Benjamini-Hochberg false discovery rate was < .05.
      The following gene sets were used in this study (shown in Table S3): 1) functional gene sets previously associated with SCZ/ASD, 2) mouse gene sets, 3) synapse gene sets from SynGO release 1.1, and 4) Gene Ontology (GO) gene sets. The functional gene sets contain synaptosome and postsynaptic density genes from Genes2Cognition, FMRP target genes (
      • Darnell J.C.
      • Van Driesche S.J.
      • Zhang C.
      • Hung K.Y.
      • Mele A.
      • Fraser C.E.
      • et al.
      FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism.
      ,
      • Ascano Jr., M.
      • Mukherjee N.
      • Bandaru P.
      • Miller J.B.
      • Nusbaum J.D.
      • Corcoran D.L.
      • et al.
      FMRP targets distinct mRNA sequence elements to regulate protein expression.
      ), and chromatin-related genes (
      • Howrigan D.P.
      • Rose S.A.
      • Samocha K.E.
      • Fromer M.
      • Cerrato F.
      • Chen W.J.
      • et al.
      Exome sequencing in schizophrenia-affected parent-offspring trios reveals risk conferred by protein-coding de novo mutations.
      ,
      • Singh T.
      • Walters J.T.R.
      • Johnstone M.
      • Curtis D.
      • Suvisaari J.
      • Torniainen M.
      • et al.
      The contribution of rare variants to risk of schizophrenia in individuals with and without intellectual disability.
      ). The mouse gene sets include 11 sets of human orthologs of mouse genes whose disruption results in neurobehavioral and nervous system abnormalities (
      • Eppig J.T.
      • Blake J.A.
      • Bult C.J.
      • Kadin J.A.
      • Richardson J.E.
      Mouse Genome Database Group
      The Mouse Genome Database (MGD): Facilitating mouse as a model for human biology and disease.
      ). The SynGO gene sets are evidence-based, expert-curated sets of synapse biology (
      • Koopmans F.
      • van Nierop P.
      • Andres-Alonso M.
      • Byrnes A.
      • Cijsouw T.
      • Coba M.P.
      • et al.
      SynGO: An evidence-based, expert-curated knowledge base for the synapse.
      ). We analyzed 59 SynGO gene sets with at least 30 genes. The GO gene sets (size 150–500 genes) were taken from the Molecular Signatures Database version 7.2 C5 collection (
      • Liberzon A.
      • Birger C.
      • Thorvaldsdottir H.
      • Ghandi M.
      • Mesirov J.P.
      • Tamayo P.
      The Molecular Signatures Database (MSigDB) hallmark gene set collection.
      ).

      Correlation of Biological Pathways

      To quantify the similarity of biological pathways among BD, SCZ, and ASD, we calculated correlation coefficients based on GO gene set results in all pairwise combinations of CNV types and disorders. To reduce the bias owing to the nonindependent nature of the gene sets, we removed gene sets that had an overlap coefficient {[size of (A intersect B)]/[size of (minimum (A, B))]} of >0.5 with regard to other gene sets, resulting in 295 gene sets for the analysis. The z score for each GO gene set was calculated from two-sided p values and odds ratios (ORs) using the following equation: z=sign(ln(OR))×|Φ1(p/2)|, where Φ1 is the inverse cumulative distribution function of the normal distribution. Therefore, the z score was positive for gene sets where the possession of CNVs increased the risk of disease and negative for gene sets where the possession of CNVs decreased the risk of disease. Pearson’s correlation coefficient among the three disorders was calculated by using the z scores. The p values for the correlation coefficients were adjusted for multiple testing using Bonferroni correction. To compare the magnitude of the two correlations, we used the R package cocor (
      • Diedenhofen B.
      • Musch J.
      cocor: A comprehensive solution for the statistical comparison of correlations.
      ), which is suitable for the comparison of coefficients calculated from two dependent groups that share a variable in common.

      CNVs in Regulatory Elements

      We examined whether case CNVs were enriched in promoters, enhancers, and topologically associating domain (TAD) boundaries in brain regions. TAD boundaries are regions bordering TADs that regulate gene expression by restricting interactions of cis-regulatory sequences to their target genes. These regulatory elements were taken from 2 sources: 1) enhancer regions in the prefrontal cortex, H3K27ac (histone H3 acetylation at lysine 27) peaks in the prefrontal, temporal, and cerebellar cortex, and TAD boundaries in the adult dorsolateral prefrontal cortex from the PsychENCODE website (http://resource.psychencode.org/), and 2) enhancer and promoter regions in 10 types of brain tissues from Reg2Map: HoneyBadger2 (https://personal.broadinstitute.org/meuleman/reg2map/HoneyBadger2_release/). H3K27ac peaks are active enhancer regions.
      For the statistical analysis, we used a logistic regression model, in which case versus control status was regressed on overlap length (one unit: 1 kb) with regulatory elements, with adjustments for array type, sex, and total length of rare CNVs. One-sided empirical p values were calculated based on 100,000 permutations, swapping case-control status. The p values were adjusted for multiple testing using Bonferroni correction.

      Results

      Identification of CNVs

      Of 8903 samples, 8708 (1818 BD cases, 3014 SCZ cases, 1205 ASD cases, and 2671 controls) passed QC (Figure 1; Table S1 in the Supplement). We obtained 25,654 rare (<1%) CNVs from all participants. The CNV characteristics are summarized in Table S4 in the Supplement. The median CNV size was 53.1 kb, and 69% and 48% were <100 kb and <50 kb, respectively. We validated 97.6% (661 of 677) of tested CNVs (Table S5 in the Supplement). For the smallest class of CNVs (10–50 kb), the validation rate was 97.0%.
      Figure thumbnail gr1
      Figure 1CNV analysis workflow. aCGH, array comparative genomic hybridization; ASD, autism spectrum disorder; BD, bipolar disorder; CNV, copy number variation; NDD, neurodevelopmental disorder; PCR, polymerase chain reaction; QC, quality control; SCZ, schizophrenia.

      Genome-wide Burden Analysis

      The results of genome-wide burden analysis are shown in Figure 2 and Table S6 in the Supplement. In BD, we found an increased burden of smaller (<100 kb) exonic deletions (OR = 1.14, pcorrected = .034). By contrast, SCZ and ASD showed the highest burden of larger (>500 kb) exonic CNVs (deletion+duplication, SCZ: OR = 1.27, ASD: OR = 1.49, pcorrected < .01).
      Figure thumbnail gr2
      Figure 2Genome-wide CNV burden. Forest plots show OR estimates and 95% confidence intervals for exonic CNV burden (CNV number) in the three disorders. Asterisks denote a significant enrichment of CNVs (pcorrected < .05). ASD, autism spectrum disorder; BD, bipolar disorder; CNV, copy number variation; OR, odds ratio; SCZ, schizophrenia.

      CNVs Linked to Neurodevelopmental Disorders

      In our sample, we identified 432 NDD-CNVs according to the American College of Medical Genetics guidelines (Figure 3A; Table S7). A significant association was found between NDD-CNVs and each disorder (BD: OR = 2.9, SCZ: OR = 3.7, ASD: OR = 4.2, pcorrected < 1 × 10−4) (Figure 3B; Table S8 in the Supplement). Figure 3C shows the percentage of individuals with each subtype of NDD-CNVs: 1) risk gene–disrupting CNVs, 2) large recurrent CNVs, 3) large nonrecurrent CNVs, and 4) sex chromosome aneuploidies. In the association analysis for BD, only the risk gene–disrupting CNVs were significant (OR = 3.6, Pcorrected = 1.0 × 10−4), whereas three subtypes were significant in SCZ and ASD (Figure 3C; Table S9 in the Supplement).
      Figure thumbnail gr3
      Figure 3NDD-CNVs. (A) Number of NDD-CNVs identified in this study. Stars indicate a significant association between the CNV and disorder (∗puncorrected < .05; ∗∗pcorrected < .05). DLG2 CNVs were significantly associated with both SCZ and ASD at pcorrected < .05. (B) Percentage of patients carrying NDD-CNVs. Frequencies of NDD-CNVs were significantly higher in each disorder compared with controls (∗∗∗pcorrected < .0001). As BD cases were analyzed by Agilent arrays only, statistical analyses for BD cases vs. controls were performed based on the data from Agilent arrays. (C) Percentage of patients carrying each subtype of NDD-CNVs: risk gene–disrupting CNVs, large recurrent CNVs, large nonrecurrent CNVs, and sex chromosome aneuploidies. ∗pcorrected < .01; ∗∗pcorrected < .001. ASD, autism spectrum disorder; BD, bipolar disorder; CNV, copy number variation; CONT, controls; del, deletion; dup, duplication; NDD, neurodevelopmental disorder; PW/AS, Prader-Willi/Angelman syndrome; RCAD synd, renal cysts and diabetes syndrome; SCZ, schizophrenia; TAR synd, thrombocytopenia–absent radius syndrome; VCFS, velocardiofacial syndrome; WBS, Williams-Beuren syndrome; XLI, X-linked ichthyosis.
      For individual NDD-CNVs, 12 showed at least nominally significant associations (Table 1; Table S10 in the Supplement). They included CNVs at three synaptic genes (DLG2, PCDH15, and ASTN2) associated with BD (Figure S1 in the Supplement). Five of 12 NDD-CNVs survived Bonferroni correction for multiple comparisons (pcorrected < .05): DLG2 CNV in BD and DLG2 CNV, 22q11.2 deletion, 1q21.1 deletion, and 47,XXX/47,XXY in SCZ.
      Table 1NDD-CNVs With at Least Nominally Significant Associations With Each Disorder
      DiagnosisNDD-CNVsFrequencyNumber of CNVsOR (95% CI)p Valuepcorrected
      CasesControlsCasesControls
      SCZDLG2 CNV0.0033010019.1 (2.4, 2460).00002.0005
      SCZ22q11.21 (VCFS region) del0.0046014021.8 (2.9, 2792).00017.0043
      SCZ1q21.1 del0.002708015.6 (1.9, 2021).00017.0043
      SCZ47,XXX/47,XXY0.0043013020.9 (2.7, 2681).00026.0065
      BDDLG2 CNV0.003306013.7 (1.6, 1789).00062.016
      ASDCNTN6 CNV0.00330.00037418.1 (1.2, 90.7).0031.078
      ASD16p11.2 dup0.00330.00037417.5 (1.3, 77.6).0034.085
      ASD22q11.21 (VCFS region) dup0.00410.00075525.8 (1.3, 34.5).0049.12
      BDPCDH15 CNV0.00280.00054513.8 (0.76, 37).019.48
      BDASTN2 CNV0.00280.00054513.8 (0.76, 37).02.5
      SCZNRXN1 CNV0.00130.00037413.5 (0.63, 35.2).0411
      SCZMACROD2 CNV0.00130.00037413.4 (0.61, 33.8).0411
      Of NDD-CNVs, 12 showed at least nominally significant associations; 5 of them were significant after Bonferroni correction for multiple comparisons (pcorrected < .05).
      ASD, autism spectrum disorder; BD, bipolar disorder; CNV, copy number variation; del, deletion; dup, duplication; NDD, neurodevelopmental disorder; OR, odds ratio; SCZ, schizophrenia; VCFS, velocardiofacial syndrome.

      Gene Set Analysis

      In 6 functional gene sets previously associated with SCZ and ASD, we found a significant enrichment of CNVs in chromatin organization and chromatin modification in BD (Figure 4A; Table S11a). In SCZ and ASD, we confirmed a significant enrichment in all 6 gene sets, except for synaptosome in ASD. In mouse gene sets and synapse gene sets, no significant enrichment was observed in BD, whereas many sets were significant in SCZ and ASD. In mouse gene sets, we found 4 significant sets common to SCZ and ASD: abnormal brain development, abnormal nervous system development, abnormal central nervous system synaptic transmission, and abnormal learning/memory/conditioning (Figure 4B; Table S11b).
      Figure thumbnail gr4
      Figure 4Results of gene set analysis (functional gene sets and mouse gene sets). (A) Functional gene sets previously associated with ASD and SCZ: synaptosome and postsynaptic density genes from Genes2Cognition; FMRP target genes from two independent datasets (
      • Darnell J.C.
      • Van Driesche S.J.
      • Zhang C.
      • Hung K.Y.
      • Mele A.
      • Fraser C.E.
      • et al.
      FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism.
      ,
      • Ascano Jr., M.
      • Mukherjee N.
      • Bandaru P.
      • Miller J.B.
      • Nusbaum J.D.
      • Corcoran D.L.
      • et al.
      FMRP targets distinct mRNA sequence elements to regulate protein expression.
      ); chromatin organization and modification genes from previous literature (
      • Howrigan D.P.
      • Rose S.A.
      • Samocha K.E.
      • Fromer M.
      • Cerrato F.
      • Chen W.J.
      • et al.
      Exome sequencing in schizophrenia-affected parent-offspring trios reveals risk conferred by protein-coding de novo mutations.
      ,
      • Singh T.
      • Walters J.T.R.
      • Johnstone M.
      • Curtis D.
      • Suvisaari J.
      • Torniainen M.
      • et al.
      The contribution of rare variants to risk of schizophrenia in individuals with and without intellectual disability.
      ). Signed log10p on the horizontal axis represents the −log10 of the p value multiplied by the sign (ln(OR)). Asterisks denote a significant enrichment of CNVs in the gene set (Benjamini-Hochberg false discovery rate < .05). (B) Mouse gene sets of human orthologs of mouse genes whose disruption causes neurobehavioral and nervous system abnormalities. Asterisks denote a significant enrichment of CNVs (Benjamini-Hochberg false discovery rate < .05). ASD, autism spectrum disorder; BD, bipolar disorder; CNV, copy number variation; SCZ, schizophrenia.
      In synapse gene sets, we identified 4 significant sets common to SCZ and ASD: process in the synapse, synaptic vesicle exocytosis, postsynaptic membrane, and integral component of postsynaptic membrane (Figure 5; Table S11c). In terms of cellular components, both presynapse and postsynapse were significant in SCZ and ASD. The gene sets with the largest effect sizes were regulation of synaptic vesicle exocytosis in SCZ and postsynaptic membrane in ASD.
      Figure thumbnail gr5
      Figure 5Results of gene set analysis (synapse gene sets: SynGO). In SCZ and ASD, significant gene sets (Benjamini-Hochberg false discovery rate < .05) are visualized in sunburst plots. Plots of bipolar disorder are omitted from the figure because no gene sets were significant. Sunburst plots are a representation of tree structures for biological processes and cellular components. Inner rings of the plot are parent terms of more specific child terms in the outer rings. Color is coded according to p values. §Significant in the analysis of deletion+duplication; †significant in the analysis of deletion; ‡significant in the analysis of duplication. ASD, autism spectrum disorder; SCZ, schizophrenia.
      In the GO gene sets, we found 1, 352, and 100 significant gene sets in BD, SCZ, and ASD, respectively (Table S11d). Among them, 81 sets were common to SCZ and ASD, and one set (covalent chromatin modification) was common to all three disorders. As shown in Figure 6A, significant gene sets were broadly classified into 14 biological pathways, 11 of which were common to SCZ and ASD.
      Figure thumbnail gr6
      Figure 6Results of GO gene set analysis and correlation analysis. (A) Nodes represent significant gene sets (Benjamini-Hochberg false discovery rate < .05) and are color-coded by diagnosis. These nodes can be broadly classified into 14 biological pathways: DNA/chromatin integrity, transcriptional regulation, cell cycle regulation, synapse/neuronal cell adhesion, transporter/channel, MAPK/ERK signaling, small GTPase signaling, Wnt signaling, cell growth/organ development, actin cytoskeleton, oxidative stress response, ubiquitin-proteasome system/autophagy, immune response, and lipid metabolism. SCZ and ASD share 11 biological pathways, particularly the DNA/chromatin integrity pathway. The purple node is GO:0016569 covalent chromatin modification, which was significant in all 3 disorders. Node size and edge thickness are proportional to the gene set size and the number of genes overlapping between gene sets, respectively. (B) Correlation of GO gene set results among the three disorders. Pairwise correlations of the z score for each GO gene set were calculated for each copy number variation type and diagnosis. The color of each box indicates the magnitude of the correlation. Correlations significantly different from zero after Bonferroni correction for all pairs of tests are marked with asterisks. ∗pcorrected < .05; ∗∗pcorrected < .0001; ∗∗∗pcorrected < .00000001. ASD, autism spectrum disorder; BD, bipolar disorder; del, deletion; deldup, deletion+duplication; dup, duplication; GO, Gene Ontology; GTPase, guanosine triphosphatase; MAPK/ERK, mitogen-activated protein kinase/extracellular signal-regulated kinase; SCZ, schizophrenia.

      Correlation of Biological Pathways

      We calculated correlations of the GO gene set results in all pairwise combinations of CNV types and disorders. In deletion+duplication, we found significant correlations among the three disorders. SCZ and ASD showed the highest degree of correlation (r = 0.48, pcorrected = 3.0 × 10−17), followed by BD and ASD (r = 0.31, pcorrected = 1.2 × 10−6), and then BD and SCZ (r = 0.25, Pcorrected = 3.7 × 10−4) (Figure 6B; Table S12 in the Supplement). The correlation coefficient between SCZ and ASD was significantly higher than that between BD and ASD (p = .0072) or between BD and SCZ (p = .0002). In deletion or duplication, we observed significant correlations among the three disorders except for the nonsignificant correlation between SCZ and BD in deletion.

      CNVs in Noncoding Regulatory Elements

      In SCZ and ASD, but not BD, CNVs were significantly (pcorrected < .05) enriched in enhancers and promoters in brain regions from HoneyBadger2 and PsychENCODE (Tables S13a and S13b in the Supplement). In most cases, deletions were significant in SCZ, whereas duplications were significant in ASD.

      Discussion

      We conducted the largest known (N = 8708) cross-disorder analysis of CNVs in BD, SCZ, and ASD. The strengths of our study are as follows: 1) the use of a high-quality and high-resolution CNV dataset (validation rate >97%, approximately 50% of CNVs <50 kb), 2) analyses of a highly homogeneous Japanese population, and 3) systematic evaluation of both genic and regulatory CNVs. Although two types of aCGH were used, we considered the batch effect to be limited for 2 reasons. First, we included array type as a covariate in all analyses for SCZ and ASD. Second, all analyses of BD cases versus controls were performed based on data from Agilent arrays.
      We found an increased burden of smaller (<100 kb) exonic deletions in BD, in contrast to the highest burden of larger (>500 kb) exonic CNVs in SCZ/ASD. Whereas an increased burden of large CNVs has been reported in SCZ/ASD (
      • Szatkiewicz J.P.
      • O’Dushlaine C.
      • Chen G.
      • Chambert K.
      • Moran J.L.
      • Neale B.M.
      • et al.
      Copy number variation in schizophrenia in Sweden.
      ,
      International Schizophrenia Consortium
      Rare chromosomal deletions and duplications increase risk of schizophrenia.
      ,
      • Pinto D.
      • Pagnamenta A.T.
      • Klei L.
      • Anney R.
      • Merico D.
      • Regan R.
      • et al.
      Functional impact of global rare copy number variation in autism spectrum disorders.
      ), the finding in BD is a novel observation and suggests that CNVs <100 kb may play an important role in BD. Interestingly, a study showed an increased burden of small (<100 kb) deletions in major depressive disorder, which was primarily in enhancer regions (
      • Zhang X.L.
      • Abdellaoui A.
      • Rucker J.
      • de Jong S.
      • Potash J.B.
      • Weissman M.M.
      • et al.
      Genome-wide burden of rare short deletions is enriched in major depressive disorder in four cohorts.
      ). Therefore, BD is similar to major depressive disorder in terms of the increased burden of small deletions, but different from major depressive disorder in terms of the direct effect on genes rather than regulatory elements.
      We found that NDD-CNVs increased the risk for BD as well as SCZ/ASD. The effect size in BD (OR = 2.9) was lower than that in SCZ/ASD (OR = 3.7–4.2), which is consistent with the notion that the contribution of CNVs to BD is smaller than that to SCZ/ASD (
      • Gordovez F.J.A.
      • McMahon F.J.
      The genetics of bipolar disorder.
      ). To our knowledge, this is the first evidence for an association between NDD-CNVs and BD. There are two reasons for this. First, the use of high-resolution aCGH enabled us to identify small CNVs, which are difficult to detect reliably with single nucleotide polymorphism arrays. Second, we carefully evaluated the pathogenicity of CNVs, including intragenic duplications according to the American College of Medical Genetics guidelines (
      • Riggs E.R.
      • Andersen E.F.
      • Cherry A.M.
      • Kantarci S.
      • Kearney H.
      • Patel A.
      • et al.
      Technical standards for the interpretation and reporting of constitutional copy-number variants: A joint consensus recommendation of the American College of Medical Genetics and Genomics (ACMG) and the Clinical Genome Resource (ClinGen).
      ,
      • Brandt T.
      • Sack L.M.
      • Arjona D.
      • Tan D.
      • Mei H.
      • Cui H.
      • et al.
      Adapting ACMG/AMP sequence variant classification guidelines for single-gene copy number variants.
      ).
      The subtype distribution of NDD-CNVs differed between BD and SCZ/ASD. Both risk gene–disrupting CNVs and large recurrent CNVs were significant in SCZ/ASD, but only the former was significant in BD. This is consistent with the results of CNV burden analysis, as the majority of risk gene–disrupting CNVs were smaller (<100 kb) exonic deletions. This also implies that risk gene–disrupting CNVs point to strong risk genes for BD as well as SCZ/ASD, as described below.
      We found a nominally significant association of 12 NDD-CNVs, 5 of which survived correction for multiple testing: CNVs at DLG2 in SCZ and BD and 22q11.21 deletion, 1q21.1 deletion, and 47,XXX/47,XXY in SCZ. Deletions at DLG2 were previously associated with SCZ and BD (
      • Kirov G.
      • Pocklington A.J.
      • Holmans P.
      • Ivanov D.
      • Ikeda M.
      • Ruderfer D.
      • et al.
      De novo CNV analysis implicates specific abnormalities of postsynaptic signalling complexes in the pathogenesis of schizophrenia.
      ,
      • Georgieva L.
      • Rees E.
      • Moran J.L.
      • Chambert K.D.
      • Milanova V.
      • Craddock N.
      • et al.
      De novo CNVs in bipolar affective disorder and schizophrenia.
      ). DLG2 plays a critical role in the molecular organization of multiprotein complexes in the postsynaptic density at excitatory synapses. Moreover, 47,XXX/47,XXY has been associated with SCZ, ASD, and BD (
      • van Rijn S.
      A review of neurocognitive functioning and risk for psychopathology in sex chromosome trisomy (47,XXY, 47,XXX, 47,XYY).
      ), but we found a specific and strong association with SCZ (OR = 20.9, pcorrected = .0065). We also found a nominally significant association between BD and two synaptic genes, PCDH15 (OR = 3.8, p = .019) and ASTN2 (OR = 3.8, p = .020). PCDH15 is responsible for Usher syndrome, characterized by retinitis pigmentosa and congenital deafness. About 20% of patients with Usher syndrome also receive a diagnosis of a psychiatric disorder (
      • Domanico D.
      • Fragiotta S.
      • Cutini A.
      • Grenga P.L.
      • Vingolo E.M.
      Psychosis, mood and behavioral disorders in Usher syndrome: Review of the literature.
      ). Neurons derived from induced pluripotent stem cells of patients with BD with PCDH15 deletion exhibit abnormalities in dendrite and synapse formation (
      • Ishii T.
      • Ishikawa M.
      • Fujimori K.
      • Maeda T.
      • Kushima I.
      • Arioka Y.
      • et al.
      In vitro modeling of the bipolar disorder and schizophrenia using patient-derived induced pluripotent stem cells with copy number variations of PCDH15 and RELN.
      ). Rare CNVs at ASTN2 were identified in patients with BD, ASD, and attention-deficit hyperactivity disorder (
      • Grozeva D.
      • Kirov G.
      • Ivanov D.
      • Jones I.R.
      • Jones L.
      • Green E.K.
      • et al.
      Rare copy number variants: A point of rarity in genetic risk for bipolar disorder and schizophrenia.
      ,
      • Lionel A.C.
      • Tammimies K.
      • Vaags A.K.
      • Rosenfeld J.A.
      • Ahn J.W.
      • Merico D.
      • et al.
      Disruption of the ASTN2/TRIM32 locus at 9q33.1 is a risk factor in males for autism spectrum disorders, ADHD and other neurodevelopmental phenotypes.
      ). All of the deletions identified in patients with BD in our study affected multiple isoforms of ASTN2. ASTN2 plays an important role in the modulation of synaptic strength by the trafficking and degradation of synaptic proteins (
      • Behesti H.
      • Fore T.R.
      • Wu P.
      • Horn Z.
      • Leppert M.
      • Hull C.
      • et al.
      ASTN2 modulates synaptic strength by trafficking and degradation of surface proteins.
      ). Taken together, these findings suggest that synaptic dysfunction is of pathogenic relevance to BD.
      The results of gene set analysis implicated chromatin modification and organization in BD pathogenesis. Consistent with this, changes in histone modification and DNA methylation were detected in postmortem brain tissue from patients with BD (
      • Ludwig B.
      • Dwivedi Y.
      Dissecting bipolar disorder complexity through epigenomic approach.
      ,
      • Bundo M.
      • Ueda J.
      • Nakachi Y.
      • Kasai K.
      • Kato T.
      • Iwamoto K.
      Decreased DNA methylation at promoters and gene-specific neuronal hypermethylation in the prefrontal cortex of patients with bipolar disorder.
      ). The involvement of chromatin modification is suggested based on the clinical efficacy of the mood stabilizer valproic acid. Valproic acid, a histone deacetylation inhibitor, causes chromatin remodeling and gene expression change (
      • Graff J.
      • Tsai L.H.
      Histone acetylation: Molecular mnemonics on the chromatin.
      ).
      In synapse gene sets, both presynapse and postsynapse were associated with the pathogenesis of SCZ and ASD. In terms of presynapse, synaptic vesicle exocytosis was implicated in both disorders. This process is essential for the maintenance of neurotransmission, and its dysregulation in SCZ/ASD has been suggested in studies of human brain tissue and animal models (
      • Egbujo C.N.
      • Sinclair D.
      • Hahn C.G.
      Dysregulations of synaptic vesicle trafficking in schizophrenia.
      ,
      • Waites C.L.
      • Garner C.C.
      Presynaptic function in health and disease.
      ). In terms of postsynapse, postsynaptic membrane was significant in SCZ/ASD. Alterations of postsynaptic membrane proteins are supported by genetic findings of SCZ/ASD–associated genes (e.g., NLGN, GPHN) and proteome analysis of brain tissue from patients (
      • Chen J.
      • Yu S.
      • Fu Y.
      • Li X.
      Synaptic proteins and receptors defects in autism spectrum disorders.
      ,
      • MacDonald M.L.
      • Garver M.
      • Newman J.
      • Sun Z.
      • Kannarkat J.
      • Salisbury R.
      • et al.
      Synaptic proteome alterations in the primary auditory cortex of individuals with schizophrenia.
      ).
      GO gene set analysis replicated previous findings that SCZ and ASD involve more extensive and similar biological pathways (
      • Kushima I.
      • Aleksic B.
      • Nakatochi M.
      • Shimamura T.
      • Okada T.
      • Uno Y.
      • et al.
      Comparative analyses of copy-number variation in autism spectrum disorder and schizophrenia reveal etiological overlap and biological insights.
      ,
      • Guilmatre A.
      • Dubourg C.
      • Mosca A.L.
      • Legallic S.
      • Goldenberg A.
      • Drouin-Garraud V.
      • et al.
      Recurrent rearrangements in synaptic and neurodevelopmental genes and shared biologic pathways in schizophrenia, autism, and mental retardation.
      ) (Figure 6A). Among others, substantial overlap was seen in the DNA/chromatin integrity pathway, which includes DNA replication, repair, recombination, and chromatin biology. Experimental evidence supports that dysregulation of these pathways can causally contribute to the pathogenesis (
      • Wang M.
      • Wei P.C.
      • Lim C.K.
      • Gallina I.S.
      • Marshall S.
      • Marchetto M.C.
      • et al.
      Increased neural progenitor proliferation in a hiPSC model of autism induces replication stress-associated genome instability.
      ,
      • Markkanen E.
      • Meyer U.
      • Dianov G.L.
      DNA damage and repair in schizophrenia and autism: Implications for cancer comorbidity and beyond.
      ,
      • McConnell M.J.
      • Moran J.V.
      • Abyzov A.
      • Akbarian S.
      • Bae T.
      • Cortes-Ciriano I.
      • et al.
      Intersection of diverse neuronal genomes and neuropsychiatric disease: The Brain Somatic Mosaicism Network.
      ). The defects of the DNA/chromatin integrity pathway may also underlie an increased genome-wide burden of rare or de novo variants in these disorders.
      Correlation analysis of gene set results showed not only strong pathway similarities (r = 0.48) between SCZ and ASD, but also weak but significant similarities (r = 0.25–0.31) between BD and SCZ/ASD (Figure 6B). This provides evidence for a shared genetic basis among these disorders, which is consistent with findings from epidemiological studies (
      • Sullivan P.F.
      • Magnusson C.
      • Reichenberg A.
      • Boman M.
      • Dalman C.
      • Davidson M.
      • et al.
      Family history of schizophrenia and bipolar disorder as risk factors for autism.
      ,
      • Song J.
      • Bergen S.E.
      • Kuja-Halkola R.
      • Larsson H.
      • Landen M.
      • Lichtenstein P.
      Bipolar disorder and its relation to major psychiatric disorders: A family-based study in the Swedish population.
      ,
      • Stahlberg O.
      • Soderstrom H.
      • Rastam M.
      • Gillberg C.
      Bipolar disorder, schizophrenia, and other psychotic disorders in adults with childhood onset AD/HD and/or autism spectrum disorders.
      ). Analyses of genome-wide association study data have reported a high genetic correlation (r = 0.7) between SCZ and BD, but a small correlation between ASD and SCZ (r = 0.21) or between ASD and BD (r = 0.18) (
      • Stahl E.A.
      • Breen G.
      • Forstner A.J.
      • McQuillin A.
      • Ripke S.
      • Trubetskoy V.
      • et al.
      Genome-wide association study identifies 30 loci associated with bipolar disorder.
      ,
      • Grove J.
      • Ripke S.
      • Als T.D.
      • Mattheisen M.
      • Walters R.K.
      • Won H.
      • et al.
      Identification of common genetic risk variants for autism spectrum disorder.
      ). While the calculation method for correlation differs from that in the present study, it is possible that the cross-disorder effects of common variants (single nucleotide polymorphisms) and rare variants (rare CNVs) may be different. Common variants may have strong cross-disorder effects on SCZ and BD, whereas rare variants may have strong cross-disorder effects on SCZ and ASD.
      In SCZ and ASD, CNVs were significantly enriched in enhancers and promoters in brain tissue. As CNVs in these noncoding regulatory elements affect gene expression (
      • Han L.
      • Zhao X.
      • Benton M.L.
      • Perumal T.
      • Collins R.L.
      • Hoffman G.E.
      • et al.
      Functional annotation of rare structural variation in the human brain.
      ), they may be implicated in risk through the dysregulation of brain-expressed genes. Previous studies have reported that variants in these regulatory elements play a role in the risk for psychiatric disorders (
      • Zhang X.L.
      • Abdellaoui A.
      • Rucker J.
      • de Jong S.
      • Potash J.B.
      • Weissman M.M.
      • et al.
      Genome-wide burden of rare short deletions is enriched in major depressive disorder in four cohorts.
      ,
      • An J.Y.
      • Lin K.
      • Zhu L.
      • Werling D.M.
      • Dong S.
      • Brand H.
      • et al.
      Genome-wide de novo risk score implicates promoter variation in autism spectrum disorder.
      ,
      • Short P.J.
      • McRae J.F.
      • Gallone G.
      • Sifrim A.
      • Won H.
      • Geschwind D.H.
      • et al.
      De novo mutations in regulatory elements in neurodevelopmental disorders.
      ). With some exceptions, deletions in SCZ and duplications in ASD were enriched in regulatory elements (Tables S13a and S13b in the Supplement), which suggests that the effect of these CNVs on gene expression may be reversed in both disorders. While there is strong evidence that SCZ and ASD share genetic commonality, there is also evidence that they have opposite genetic bases (e.g., SCZ is associated with 22q11.2 deletion, whereas ASD is associated with 22q11.2 duplication) (
      • Crespi B.
      • Stead P.
      • Elliot M.
      Comparative genomics of autism and schizophrenia.
      ).
      In conclusion, BD and SCZ/ASD differ in terms of CNV burden, characteristics of NDD-CNVs, and regulatory CNVs. On the other hand, they have shared molecular mechanisms, including chromatin biology. The BD risk genes identified in the present study could provide insight into the pathogenesis of BD.

      Acknowledgments and Disclosures

      This work was supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT) and the Ministry of Health, Labour and Welfare of Japan , Japan Agency for Medical Research and Development (Grant Nos. JP20dm0107087 [to NOz] , JP21wm0425007 [to NOz] , JP21dm0207075 [to NOz] , JP21ak0101113 [to NOz] , JP20dk0307075 [to NOz, TAK] , JP20dk0307081 [to NOz] , JP21dk0307103 [to NOz, RH] , JP21km0405216 [to MN, IK] , JP21ek0109411 [to IK] , JP20dm0107160 [to TOka] , JP18dm0107088 [to MIt] , JP19dm0107088 [to MIt] , JP20dm0107088 [to MIt] , JP20dm0107092 [to KY] , JP21dm0207069 [to MY] , JP21wm0425019 [to YK] , JP21dm0207074 [to YK] , JP19lk0201071 [to HYamas] , JP20lk0201116 to [HYamas] , JP16dm0107134 [to HYamas] , JP16dk0307029 [to MS] , JP20dm0107083 [to TY] , JP20km0405208 [to MIk, TK] , JP20dm0207074 [to TK] , JP20dm0107097 [to NI, MIk] , JP21wm0425008 [to NI, MIk] , JP21tm0424220 [to MIk] , JP20km0405201 [to NI, MIk], and JP21wm0525024 [to MIk] ), Japan Society for the Promotion of Science (KAKENHI Grant Nos. JP23110506 [to NOz] , JP23700443 [to NOz] , JP25110715 [to NOz] , JP25460284 [to NOz] , JP17H05090 [to IK] , JP15K19720 [to IK] , JP18H04040 [to NOz] , JP21K07543 [to IK] , JP20K20602 [to NOz] , JP21H00194 [to IK] , JP21H04815 [to NOz] , JP17K10295 [to MY] , JP21K07498 [to MY] , JP19K17061 [to SS] , JP18K07550 [to TTa] , JP16H05380 [to MA] , JP17H05930 [to MA] , JP19H04887 [to MA] , JP20H03608 [to MA] , JP21H00180 [to YK] , JP19K08053 [to YK] , JP20H03598 [to MS] , JP18H05435 [to TK] , JP18H05428 [to TK] , JP19K08081 [to KO] , JP17H04251 [to MIk] , JP16H06277 [to MN] , and JP21H02854 [to MIk] ), Private University Research Branding Project from MEXT [to NI], Collaborative Research Project of the Brain Research Institute, Niigata University (Grant No. 201917) [to YK], National Center of Neurology and Psychiatry Intramural Research Grant (3-1) for Neurological and Psychiatric Disorders [to RH], Uehara Memorial Foundation [to MA, IK], SENSHIN Medical Research Foundation [to NI, MIk, IK, MA], and Sumitomo Foundation [to MA].
      We thank all patients and their families for participating in this study. We also thank Mami Yoshida, Kiyori Monta, Hiromi Noma, and Yukari Mitsui for their technical assistance.
      IK has received research/grant support from AMED, MEXT/JSPS, Novartis Pharma, GlaxoSmithKline, Takeda Pharmaceutical Co., Ltd., Hori Sciences and Arts Foundation, Uehara Memorial Foundation, and the SENSHIN Medical Research Foundation. SIs has received personal fees from Janssen Pharmaceutical, Sumitomo Dainippon Pharma Co., Ltd., Eisai Co., Ltd., and Meiji Seika Pharma and research/grant support from Eli Lilly. SNu has received research grants, rewards for lectures, and donations from Astellas Pharma Inc., Eisai Co., Ltd., Otsuka Pharmaceutical Co. Ltd., Sumitomo Dainippon Pharma Co., Ltd., Eli Lilly Japan K.K., Novartis Pharma K.K., Pfizer Inc., Meiji Seika Pharma Co. Ltd., Mochida Pharmaceutical Co., Ltd., Janssen Pharmaceutical K.K., Kyowa Pharmaceutical Industry Co., Ltd., Takeda Pharmaceutical Co., Ltd., and Yoshitomiyakuhin Co. MIt has been awarded patents regarding the therapeutic use of pyridoxamine for schizophrenia. NOz has received research support or speakers’ honoraria from, or has served as a consultant to, Sumitomo Dainippon Pharma Co., Ltd., Eisai Co., Ltd., Otsuka, KAITEKI, Mitsubishi Tanabe, Shionogi, Eli Lilly, Mochida, DAIICHI SANKYO, Nihon Medi-Physics, Takeda, Meiji Seika Pharma, EA Pharma, Pfizer, MSD, Lundbeck Japan, and Taisho Pharma Co. outside the submitted work. NI has received research support or speakers’ honoraria from, or has served as a consultant to, Jansen, GlaxoSmithKline, Eli Lilly, Otsuka, Shionogi, Sumitomo Dainippon Pharma Co., Ltd., Tanabe Mitsubishi, and Daiichi-Sankyo. TK has received grants and personal fees from AMED and MEXT/JSPS during the conduct of the study and personal fees from Kyowa Hakko Kirin Co., Ltd., Eli Lilly Japan K.K., GlaxoSmithKline, Taisho Pharma Co., Meiji Seika Pharma Co., Ltd., Pfizer Japan Inc., Mochida Pharmaceutical Co., Ltd., Janssen Pharmaceutical K.K., Janssen Asia Pacific, Yoshitomiyakuhin Co., Astellas Pharma Inc., Nippon Boehringer Ingelheim Co., Ltd., MSD K.K., Kyowa Pharmaceutical Industry Co., Ltd., Taisho Pharmaceutical Co., Ltd., and Taisho Toyama Pharmaceutical Co., Ltd.; grants and personal fees from Otsuka Pharmaceutical Co., Ltd., Sumitomo Dainippon Pharma Co., Ltd., Shionogi & Co., Ltd., Takeda Pharmaceutical Co., Ltd., Eisai Co., Ltd., and Mitsubishi Tanabe Pharma Corp.; and grants from Teijin Pharma outside the submitted work. All other authors report no biomedical financial interests or potential conflicts of interest.

      References

        • Sullivan P.F.
        • Magnusson C.
        • Reichenberg A.
        • Boman M.
        • Dalman C.
        • Davidson M.
        • et al.
        Family history of schizophrenia and bipolar disorder as risk factors for autism.
        Arch Gen Psychiatry. 2012; 69: 1099-1103
        • Song J.
        • Bergen S.E.
        • Kuja-Halkola R.
        • Larsson H.
        • Landen M.
        • Lichtenstein P.
        Bipolar disorder and its relation to major psychiatric disorders: A family-based study in the Swedish population.
        Bipolar Disord. 2015; 17: 184-193
        • Stahlberg O.
        • Soderstrom H.
        • Rastam M.
        • Gillberg C.
        Bipolar disorder, schizophrenia, and other psychotic disorders in adults with childhood onset AD/HD and/or autism spectrum disorders.
        J Neural Transm. 2004; 111: 891-902
        • Pinto D.
        • Delaby E.
        • Merico D.
        • Barbosa M.
        • Merikangas A.
        • Klei L.
        • et al.
        Convergence of genes and cellular pathways dysregulated in autism spectrum disorders.
        Am J Hum Genet. 2014; 94: 677-694
        • Marshall C.R.
        • Howrigan D.P.
        • Merico D.
        • Thiruvahindrapuram B.
        • Wu W.
        • Greer D.S.
        • et al.
        Contribution of copy number variants to schizophrenia from a genome-wide study of 41,321 subjects.
        Nat Genet. 2017; 49: 27-35
        • Szatkiewicz J.P.
        • O’Dushlaine C.
        • Chen G.
        • Chambert K.
        • Moran J.L.
        • Neale B.M.
        • et al.
        Copy number variation in schizophrenia in Sweden.
        Mol Psychiatry. 2014; 19: 762-773
        • Sebat J.
        • Lakshmi B.
        • Malhotra D.
        • Troge J.
        • Lese-Martin C.
        • Walsh T.
        • et al.
        Strong association of de novo copy number mutations with autism.
        Science. 2007; 316: 445-449
        • Xu B.
        • Roos J.L.
        • Levy S.
        • van Rensburg E.J.
        • Gogos J.A.
        • Karayiorgou M.
        Strong association of de novo copy number mutations with sporadic schizophrenia.
        Nat Genet. 2008; 40: 880-885
        • Kushima I.
        • Aleksic B.
        • Nakatochi M.
        • Shimamura T.
        • Okada T.
        • Uno Y.
        • et al.
        Comparative analyses of copy-number variation in autism spectrum disorder and schizophrenia reveal etiological overlap and biological insights.
        Cell Rep. 2018; 24: 2838-2856
        • Nakatochi M.
        • Kushima I.
        • Ozaki N.
        Implications of germline copy-number variations in psychiatric disorders: Review of large-scale genetic studies.
        J Hum Genet. 2021; 66: 25-37
        • Kirov G.
        • Pocklington A.J.
        • Holmans P.
        • Ivanov D.
        • Ikeda M.
        • Ruderfer D.
        • et al.
        De novo CNV analysis implicates specific abnormalities of postsynaptic signalling complexes in the pathogenesis of schizophrenia.
        Mol Psychiatry. 2012; 17: 142-153
        • Green E.K.
        • Rees E.
        • Walters J.T.
        • Smith K.G.
        • Forty L.
        • Grozeva D.
        • et al.
        Copy number variation in bipolar disorder.
        Mol Psychiatry. 2016; 21: 89-93
        • Charney A.W.
        • Stahl E.A.
        • Green E.K.
        • Chen C.Y.
        • Moran J.L.
        • Chambert K.
        • et al.
        Contribution of rare copy number variants to bipolar disorder risk is limited to schizoaffective cases.
        Biol Psychiatry. 2019; 86: 110-119
        • Malhotra D.
        • McCarthy S.
        • Michaelson J.J.
        • Vacic V.
        • Burdick K.E.
        • Yoon S.
        • et al.
        High frequencies of de novo CNVs in bipolar disorder and schizophrenia.
        Neuron. 2011; 72: 951-963
        • Priebe L.
        • Degenhardt F.A.
        • Herms S.
        • Haenisch B.
        • Mattheisen M.
        • Nieratschker V.
        • et al.
        Genome-wide survey implicates the influence of copy number variants (CNVs) in the development of early-onset bipolar disorder.
        Mol Psychiatry. 2012; 17: 421-432
        • Gordovez F.J.A.
        • McMahon F.J.
        The genetics of bipolar disorder.
        Mol Psychiatry. 2020; 25: 544-559
        • Grozeva D.
        • Kirov G.
        • Ivanov D.
        • Jones I.R.
        • Jones L.
        • Green E.K.
        • et al.
        Rare copy number variants: A point of rarity in genetic risk for bipolar disorder and schizophrenia.
        Arch Gen Psychiatry. 2010; 67: 318-327
        • Martin C.L.
        • Wain K.E.
        • Oetjens M.T.
        • Tolwinski K.
        • Palen E.
        • Hare-Harris A.
        • et al.
        Identification of neuropsychiatric copy number variants in a health care system population.
        JAMA Psychiatry. 2020; 77: 1276-1285
        • Yang Z.
        • Wu H.
        • Lee P.H.
        • Tsetsos F.
        • Davis L.K.
        • Yu D.
        • et al.
        Investigating shared genetic basis across Tourette syndrome and comorbid neurodevelopmental disorders along the impulsivity-compulsivity spectrum.
        Biol Psychiatry. 2021; 90: 317-327
        • Zarrei M.
        • Burton C.L.
        • Engchuan W.
        • Young E.J.
        • Higginbotham E.J.
        • MacDonald J.R.
        • et al.
        A large data resource of genomic copy number variation across neurodevelopmental disorders.
        NPJ Genom Med. 2019; 4: 26
        • McGrath L.M.
        • Yu D.
        • Marshall C.
        • Davis L.K.
        • Thiruvahindrapuram B.
        • Li B.
        • et al.
        Copy number variation in obsessive-compulsive disorder and Tourette syndrome: A cross-disorder study.
        J Am Acad Child Adolesc Psychiatry. 2014; 53: 910-919
        • Zhang F.
        • Lupski J.R.
        Non-coding genetic variants in human disease.
        Hum Mol Genet. 2015; 24: R102-R110
        • Murakawa Y.
        • Yoshihara M.
        • Kawaji H.
        • Nishikawa M.
        • Zayed H.
        • Suzuki H.
        • et al.
        Enhanced identification of transcriptional enhancers provides mechanistic insights into diseases.
        Trends Genet. 2016; 32: 76-88
        • Spielmann M.
        • Klopocki E.
        CNVs of noncoding cis-regulatory elements in human disease.
        Curr Opin Genet Dev. 2013; 23: 249-256
        • ENCODE Project Consortium
        An integrated encyclopedia of DNA elements in the human genome.
        Nature. 2012; 489: 57-74
        • Kundaje A.
        • Meuleman W.
        • Ernst J.
        • Bilenky M.
        • Yen A.
        • et al.
        • Roadmap Epigenomics Consortium
        Integrative analysis of 111 reference human epigenomes.
        Nature. 2015; 518: 317-330
        • Lepretre F.
        • Villenet C.
        • Quief S.
        • Nibourel O.
        • Jacquemin C.
        • Troussard X.
        • et al.
        Waved aCGH: To smooth or not to smooth.
        Nucleic Acids Res. 2010; 38: e94
        • Kushima I.
        • Aleksic B.
        • Nakatochi M.
        • Shimamura T.
        • Shiino T.
        • Yoshimi A.
        • et al.
        High-resolution copy number variation analysis of schizophrenia in Japan.
        Mol Psychiatry. 2017; 22: 430-440
        • Gonzalez-Mantilla A.J.
        • Moreno-De-Luca A.
        • Ledbetter D.H.
        • Martin C.L.
        A Cross-disorder method to identify novel candidate genes for developmental brain disorders.
        JAMA Psychiatry. 2016; 73: 275-283
        • Zhao G.H.
        • Li K.K.
        • Li B.
        • Wang Z.
        • Fang Z.H.
        • Wang X.M.
        • et al.
        Gene4Denovo: An integrated database and analytic platform for de novo mutations in humans.
        Nucleic Acids Res. 2020; 48: D913-D926
        • Turner T.N.
        • Yi Q.
        • Krumm N.
        • Huddleston J.
        • Hoekzema K.
        • F Stessman H.A.
        • et al.
        denovo-db: A compendium of human de novo variants.
        Nucleic Acids Res. 2017; 45: D804-D811
        • Abrahams B.S.
        • Arking D.E.
        • Campbell D.B.
        • Mefford H.C.
        • Morrow E.M.
        • Weiss L.A.
        • et al.
        SFARI Gene 2.0: A community-driven knowledgebase for the autism spectrum disorders (ASDs).
        Mol Autism. 2013; 4: 36
        • Riggs E.R.
        • Andersen E.F.
        • Cherry A.M.
        • Kantarci S.
        • Kearney H.
        • Patel A.
        • et al.
        Technical standards for the interpretation and reporting of constitutional copy-number variants: A joint consensus recommendation of the American College of Medical Genetics and Genomics (ACMG) and the Clinical Genome Resource (ClinGen).
        Genet Med. 2020; 22: 245-257
        • Brandt T.
        • Sack L.M.
        • Arjona D.
        • Tan D.
        • Mei H.
        • Cui H.
        • et al.
        Adapting ACMG/AMP sequence variant classification guidelines for single-gene copy number variants.
        Genet Med. 2020; 22: 336-344
        • Raychaudhuri S.
        • Korn J.M.
        • McCarroll S.A.
        • Altshuler D.
        • Sklar P.
        • et al.
        • International Schizophrenia Consortium
        Accurately assessing the risk of schizophrenia conferred by rare copy-number variation affecting genes with brain function.
        PLoS Genet. 2010; 6e1001097
        • Benjamini Y.
        • Hochberg Y.
        Controlling the false discovery rate—a practical and powerful approach to multiple testing.
        J R Stat Soc B. 1995; 57: 289-300
        • Darnell J.C.
        • Van Driesche S.J.
        • Zhang C.
        • Hung K.Y.
        • Mele A.
        • Fraser C.E.
        • et al.
        FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism.
        Cell. 2011; 146: 247-261
        • Ascano Jr., M.
        • Mukherjee N.
        • Bandaru P.
        • Miller J.B.
        • Nusbaum J.D.
        • Corcoran D.L.
        • et al.
        FMRP targets distinct mRNA sequence elements to regulate protein expression.
        Nature. 2012; 492: 382-386
        • Howrigan D.P.
        • Rose S.A.
        • Samocha K.E.
        • Fromer M.
        • Cerrato F.
        • Chen W.J.
        • et al.
        Exome sequencing in schizophrenia-affected parent-offspring trios reveals risk conferred by protein-coding de novo mutations.
        Nat Neurosci. 2020; 23: 185-193
        • Singh T.
        • Walters J.T.R.
        • Johnstone M.
        • Curtis D.
        • Suvisaari J.
        • Torniainen M.
        • et al.
        The contribution of rare variants to risk of schizophrenia in individuals with and without intellectual disability.
        Nat Genet. 2017; 49: 1167-1173
        • Eppig J.T.
        • Blake J.A.
        • Bult C.J.
        • Kadin J.A.
        • Richardson J.E.
        • Mouse Genome Database Group
        The Mouse Genome Database (MGD): Facilitating mouse as a model for human biology and disease.
        Nucleic Acids Res. 2015; 43: D726-D736
        • Koopmans F.
        • van Nierop P.
        • Andres-Alonso M.
        • Byrnes A.
        • Cijsouw T.
        • Coba M.P.
        • et al.
        SynGO: An evidence-based, expert-curated knowledge base for the synapse.
        Neuron. 2019; 103: 217-234.e214
        • Liberzon A.
        • Birger C.
        • Thorvaldsdottir H.
        • Ghandi M.
        • Mesirov J.P.
        • Tamayo P.
        The Molecular Signatures Database (MSigDB) hallmark gene set collection.
        Cell Syst. 2015; 1: 417-425
        • Diedenhofen B.
        • Musch J.
        cocor: A comprehensive solution for the statistical comparison of correlations.
        PLoS One. 2015; 10e0121945
        • International Schizophrenia Consortium
        Rare chromosomal deletions and duplications increase risk of schizophrenia.
        Nature. 2008; 455: 237-241
        • Pinto D.
        • Pagnamenta A.T.
        • Klei L.
        • Anney R.
        • Merico D.
        • Regan R.
        • et al.
        Functional impact of global rare copy number variation in autism spectrum disorders.
        Nature. 2010; 466: 368-372
        • Zhang X.L.
        • Abdellaoui A.
        • Rucker J.
        • de Jong S.
        • Potash J.B.
        • Weissman M.M.
        • et al.
        Genome-wide burden of rare short deletions is enriched in major depressive disorder in four cohorts.
        Biol Psychiatry. 2019; 85: 1065-1073
        • Georgieva L.
        • Rees E.
        • Moran J.L.
        • Chambert K.D.
        • Milanova V.
        • Craddock N.
        • et al.
        De novo CNVs in bipolar affective disorder and schizophrenia.
        Hum Mol Genet. 2014; 23: 6677-6683
        • van Rijn S.
        A review of neurocognitive functioning and risk for psychopathology in sex chromosome trisomy (47,XXY, 47,XXX, 47,XYY).
        Curr Opin Psychiatry. 2019; 32: 79-84
        • Domanico D.
        • Fragiotta S.
        • Cutini A.
        • Grenga P.L.
        • Vingolo E.M.
        Psychosis, mood and behavioral disorders in Usher syndrome: Review of the literature.
        Med Hypothesis Discov Innov Ophthalmol. 2015; 4: 50-55
        • Ishii T.
        • Ishikawa M.
        • Fujimori K.
        • Maeda T.
        • Kushima I.
        • Arioka Y.
        • et al.
        In vitro modeling of the bipolar disorder and schizophrenia using patient-derived induced pluripotent stem cells with copy number variations of PCDH15 and RELN.
        eNeuro. 2019; (6:ENEURO.0403-18.2019)
        • Lionel A.C.
        • Tammimies K.
        • Vaags A.K.
        • Rosenfeld J.A.
        • Ahn J.W.
        • Merico D.
        • et al.
        Disruption of the ASTN2/TRIM32 locus at 9q33.1 is a risk factor in males for autism spectrum disorders, ADHD and other neurodevelopmental phenotypes.
        Hum Mol Genet. 2014; 23: 2752-2768
        • Behesti H.
        • Fore T.R.
        • Wu P.
        • Horn Z.
        • Leppert M.
        • Hull C.
        • et al.
        ASTN2 modulates synaptic strength by trafficking and degradation of surface proteins.
        Proc Natl Acad Sci U S A. 2018; 115: E9717-E9726
        • Ludwig B.
        • Dwivedi Y.
        Dissecting bipolar disorder complexity through epigenomic approach.
        Mol Psychiatry. 2016; 21: 1490-1498
        • Bundo M.
        • Ueda J.
        • Nakachi Y.
        • Kasai K.
        • Kato T.
        • Iwamoto K.
        Decreased DNA methylation at promoters and gene-specific neuronal hypermethylation in the prefrontal cortex of patients with bipolar disorder.
        Mol Psychiatry. 2021; 26: 3407-3418
        • Graff J.
        • Tsai L.H.
        Histone acetylation: Molecular mnemonics on the chromatin.
        Nat Rev Neurosci. 2013; 14: 97-111
        • Egbujo C.N.
        • Sinclair D.
        • Hahn C.G.
        Dysregulations of synaptic vesicle trafficking in schizophrenia.
        Curr Psychiatry Rep. 2016; 18: 77
        • Waites C.L.
        • Garner C.C.
        Presynaptic function in health and disease.
        Trends Neurosci. 2011; 34: 326-337
        • Chen J.
        • Yu S.
        • Fu Y.
        • Li X.
        Synaptic proteins and receptors defects in autism spectrum disorders.
        Front Cell Neurosci. 2014; 8: 276
        • MacDonald M.L.
        • Garver M.
        • Newman J.
        • Sun Z.
        • Kannarkat J.
        • Salisbury R.
        • et al.
        Synaptic proteome alterations in the primary auditory cortex of individuals with schizophrenia.
        JAMA Psychiatry. 2020; 77: 86-95
        • Guilmatre A.
        • Dubourg C.
        • Mosca A.L.
        • Legallic S.
        • Goldenberg A.
        • Drouin-Garraud V.
        • et al.
        Recurrent rearrangements in synaptic and neurodevelopmental genes and shared biologic pathways in schizophrenia, autism, and mental retardation.
        Arch Gen Psychiatry. 2009; 66: 947-956
        • Wang M.
        • Wei P.C.
        • Lim C.K.
        • Gallina I.S.
        • Marshall S.
        • Marchetto M.C.
        • et al.
        Increased neural progenitor proliferation in a hiPSC model of autism induces replication stress-associated genome instability.
        Cell Stem Cell. 2020; 26: 221-233.e226
        • Markkanen E.
        • Meyer U.
        • Dianov G.L.
        DNA damage and repair in schizophrenia and autism: Implications for cancer comorbidity and beyond.
        Int J Mol Sci. 2016; 17: 856
        • McConnell M.J.
        • Moran J.V.
        • Abyzov A.
        • Akbarian S.
        • Bae T.
        • Cortes-Ciriano I.
        • et al.
        Intersection of diverse neuronal genomes and neuropsychiatric disease: The Brain Somatic Mosaicism Network.
        Science. 2017; 356eaal1641
        • Stahl E.A.
        • Breen G.
        • Forstner A.J.
        • McQuillin A.
        • Ripke S.
        • Trubetskoy V.
        • et al.
        Genome-wide association study identifies 30 loci associated with bipolar disorder.
        Nat Genet. 2019; 51: 793-803
        • Grove J.
        • Ripke S.
        • Als T.D.
        • Mattheisen M.
        • Walters R.K.
        • Won H.
        • et al.
        Identification of common genetic risk variants for autism spectrum disorder.
        Nat Genet. 2019; 51: 431-444
        • Han L.
        • Zhao X.
        • Benton M.L.
        • Perumal T.
        • Collins R.L.
        • Hoffman G.E.
        • et al.
        Functional annotation of rare structural variation in the human brain.
        Nat Commun. 2020; 11: 2990
        • An J.Y.
        • Lin K.
        • Zhu L.
        • Werling D.M.
        • Dong S.
        • Brand H.
        • et al.
        Genome-wide de novo risk score implicates promoter variation in autism spectrum disorder.
        Science. 2018; 362eaat6576
        • Short P.J.
        • McRae J.F.
        • Gallone G.
        • Sifrim A.
        • Won H.
        • Geschwind D.H.
        • et al.
        De novo mutations in regulatory elements in neurodevelopmental disorders.
        Nature. 2018; 555: 611-616
        • Crespi B.
        • Stead P.
        • Elliot M.
        Comparative genomics of autism and schizophrenia.
        Proc Natl Acad Sci U S A. 2010; 107: 1736-1741