Study Report

Reference	Saito T., 2014 PMID: 24888570
Citation	Saito, T., et al. (2014). "Replication and cross-phenotype study based upon schizophrenia GWASs data in the Japanese population: Support for association of MHC region with psychosis." Am J Med Genet B Neuropsychiatr Genet.
Disease Type	Bipolar Disorder & Schizophrenia
Study Design	case-control
Study Type	Genome-wide association study
Sample Size	1032 SCZ, 1012 BD(BD typeI=621, BD typeII=380, schizoaffective disorder [SA]=7, BD type information not available=4) and 993 controls.
SNP/Region/Marker Size	51 SNPs
Predominant Ethnicity	Mongloid
Population	Japanese
Gender	The first-set: BD: 51.8% female subjects of 1,012 patients; SZ: 48.3% female subjects of 1,032 patients; Controls: 51.1% female subjects of 993 healthy controls. The replication set: BD: 54.6% female subjects of 821 patients, SD=14.4; SZ: 45.1% female subjects of 1,808 patients; Controls: 57.2% female subjects of 2,321 healthy controls.
Age Group	adults : The first-set: BD: mean age=50.7, SD=14.3; SZ: mean age=46.8, SD=14.8; Controls: mean age=49.7, SD=14.0. The replication set: BD: mean age=48.2, SD=14.4; SZ: mean age=49.8, SD=14.8; Controls: mean age=42.3, SD=14.2.

Sample Diagnosis	DSM-IV
Sample Status	All subjects were unrelated to each other and ethnically Japanese. Patients with bipolar disorder (BD) were collected at each university participating in the Collaborative Study of Mood Disorder (COSMO). The patients with schizophrenia (SCZ) were collected at Fujita Health University and Okayama University. The patients were diagnosed according to DSM-IV criteria with the consensus of at least two experienced psychiatrists based on unstructured interviews and a review of medical records.
Replication Size	1808 SCZ, 821 BD(BD type I =387, BD type II = 344, SA=89, unknown type of diagnosis=1) and 2321 controls
Technique	All SNPs were genotyped using the Sequenom iPLEX Gold system (Sequenom, San Diego, CA).
Statistical Method	All statistical procedures were performed using PLINK version 1.07.We assessed the allelic association(two-tailed) of the selected SNPs and the three phenotypes. To account for linkage disequilibrium (LD) between SNPs selected for the analysis, we used the Single Nucleotide Polymorphism Spectral Decomposition (SNPSpD) interface to identify a sufficient number of independent variables for the first-set screening analysis (N=42). Ameta-analysis of the SNPs was conducted by combining the screening and replication datasets. A binary effects model, one of the novel random effects approaches, which is providing more statistical power under the existence of heterogeneity, was applied in each analysis using METASOFT program.
Result Summary	Nine SNPs revealed nominal association signals for all comparisons (Puncorrected<0.05), of which two SNPs located in the major histocompatibility complex region(rs7759855in zinc finger and SCAN domain containing 31 [ZSCAN31] and rs1736913 in HLA-F antisense RNA1 [HLA-F-AS1]) were further assessed in the second-set replication samples. The associations were confirmed for rs7759855 (Pcorrected=0.026 for psychosis; Pcorrected=0.032 for SCZ), although the direction of effect was opposite to that in the original GWAS of the Chinese population. Finally, a meta-analysis was conducted using our two samples and using our data and data from PsychiatricGWASConsortium (PGC), which have shown the same direction of effect. SNP in ZSCAN31(rs7759855) had the strongest association with the phenotypes(best P=6.8*10-5 for psychosis: present plus PGC results).