Genome-wide association study of schizophrenia in a Chinese population

Abstract

Objective : Schizophrenia is a severe psychiatric disease with a lifetime risk of approximately 1%. Although the heritability of schizophrenia is high, few susceptibility genes have been identified to date. Genome-wide association study (GWAS) has proved to be a powerful tool in dissecting the genetic basis of complex diseases. To date, most GWAS on schizophrenia were conducted on Caucasians. However, allele frequencies, linkage disequilibrium patterns and genetic or environmental backgrounds may differ among populations, rendering some variants more readily discovered in one population than the other. Methods : We have performed a GWAS on schizophrenia in a Chinese Han population. Samples were genotyped using the Illumina Human610-Quad BeadChip. After quality control procedures, the dataset consisted of 473931 SNPs from 481 cases and 2034 controls. Association between each SNP and disease status was tested by logistic regression under an additive model. Results : In total 76 SNPs achieved nominal p-values less than 1E-4. However, no SNPs achieved genome-wide significance (5E-8). We also extracted the association results of SNPs lying within 33 candidate genes that were implicated in previous GWAS or listed in the SZgene database. The QQ plot however shows no evidence of enrichment of significant signals. The most significant SNP in the candidate gene analysis belongs to the PLXNA2 gene (p=5.93E-4). In addition, we extracted results from SNPs within linkage hotspots suggested by a recent meta-analysis and SNPs within the MHC region. Both QQ plots showed positive deviations from the predicted line corresponding to the complete null hypothesis, providing evidence that some associated SNPs in these regions represent true positives. Conclusion: We have not been able to detect any SNPs passing genome-wide significance. Common variants with large effect sizes are unlikely to exist for schizophrenia. Larger sample sizes are required to detect susceptibility variants for the disease.