Design and Analysis of Multiple Diseases GWAS Without Controls

• Activity Number: 623
• Type: Contributed
• Date/Time: Thursday, August 2, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Zhongxue Chen*+
• Companies: The University of Texas at Houston
• Keywords: chi-square partition ; genetic association ; robust test ; trend test ; SNP

Abstract

Some genome-wide association studies (GWAS) with multiple diseases and shared controls have been conducted. If the data are appropriately analyzed, this design can have several advantages, including improved power and less money and time to generate data. We proposed a new design and analysis strategy for GWAS with multiple diseases but without controls. Through a simulation study and an application to a real data set of GWAS with multiple diseases and common controls, we show that the proposed statistical association testing method is more powerful than the test using single disease and common controls only; furthermore, it has comparable power as the overall test which uses all data including the controls. Our findings indicate the proposed design and analysis of the GWAS with multiple diseases but without controls are more efficient than the usual case control GWAS and those with shared controls.