On the meta-Analysis of Genomewide Association Studies: A Robust and Efficient Approach to Combine Population and Family-Based Studies

• Activity Number: 564
• Type: Topic Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Christoph Lange and Sungho Won*+
• Companies: Chung-Ang University and Harvard School of Public Health
• Keywords: GWA ; association study

Abstract

For the meta-analysis of genome-wide association studies, we propose a new mixed-model approach that combines population-based and family-based samples. The approach achieves the same power levels as a standard meta analysis which combines the different test statistics or p-values across studies. However, by virtue of its design, the proposed approach is completely robust against population admixture, i.e. no adjustments for population admixture are required, not even for the population-based samples. Using simulation studies, we examine the power of the proposed method and compare it to standard approaches that are applied in the meta-analysis of genome-wide association studies. The practical features of the approach are illustrated by a meta analysis of 3 genome-wide association studies for Alzheimers disease.