Compairing quantitative trait analysis to qualitative trait analysis for complex traits disease: A genome wide association study for hyperlipidemia

摘要

Current standard genome-wide association studies (GWAS) have relied on the simple analysis by focusing on the association between single genetic factor and one single common complex trait. However, since most common complex traits are associated with multiple genetic factors and their epistasis, this simple analysis is not powerful enough to detect multiple genetic factors. Furthermore, in many GWAS, one binary trait is commonly used and it is usually a summary trait derived from several quantitative traits. For example, a binary trait representing hyperlipidemia status is defined by combining four quantitative traits: Total cholesterol (Tchl), High density lipoprotein (HDL) cholesterol, Low density lipoprotein (LDL), and cholesterol and Triglycerides (TG). More information can be extracted from these quantitative traits than from one summary binary trait. However, not many methods have been proposed to account for the multiple traits simultaneously. In this study, we propose the following simple stepwise strategy to increase the power of detecting multiple genetic factors jointly for the multiple traits: (1) prescreening, (2) joint identification of putative SNPs based on elastic-net (EN) variable selection, and (3) collapsing. Joint identification of multiple genetic factors would be more powerful and provide better prediction on complex traits. We illustrated our approach with a large scale genome-wide dataset from a Korean population and identified the genetic factors associated with lipid-related traits.