Genome-wide gene-set analysis identifies different patterns of genetic sharing across complex phenotypes

Abstract

As an important complement to individual SNP analysis and gene-based analysis for a typical genome-wide association study (GWAS), GWAS gene-set analysis has the potential to discover hidden disease susceptibility genes by combining statistical evidence with biological knowledge. Recently a gene-set analytical tool ‘MAGMA’ has been developed to handle polygenic traits using more reasonable and powerful competitive test. Here we adopt both classical gene-set enrichment analysis (hypergeometric test implemented in KGG; http://statgenpro.psychiatry.hku.hk/limx/kgg/) and MAGMA on GWAS summary statistics of six diseases or traits (Crohn’s disease (CD), rheumatoid arthritis (RA), schizophrenia (SCZ), bipolar disorder (BPD), low-density cholesterol (LDL) and high-density cholesterol (HDL) level) to look for pathways/gene-sets important for their normal functioning or abnormal pathogenesis. Though no gene-set was significantly associated with two psychiatric diseases, we found a few gene-sets enriched with susceptibility genes for other four phenotypes. Interestingly those LDL/HDL shared gene-sets involving in lipid metabolism or transport are mainly due to pleiotropic apolipoprotein genes for both phenotypes; however, those CD/RA shared gene-sets are ascribed to different phenotype-specific genes which are all important to immune response. Our study reveals that genetic sharing at advanced gene-set level can sometimes provide better perspective to explain disease comorbidity.