Evaluation of Mixed-Model for Control of Population Stratification in Human Genetic Studies

• Activity Number: 132
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: ENAR
• Author(s): Nianjun Liu* and Hongyu Zhao and David B. Allison+
• Companies: The University of Alabama at Birmingham and Yale University and The University of Alabama at Birmingham
• Keywords: mixed-model ; genetics ; population stratification ; kinship

Abstract

In genetic studies, associations may be confounded by unrecognized population stratification. Recently many methods have been proposed to address this issue. Among them, the approaches based on mixed-model have several advantages. However, these approaches have not been fully evaluated on human data. The objectives of this study are to (1) evaluate various methods based on mixed-model for genetic association mapping using human data, (2) compare the performance of different marker-based kinship matrices, (3) understand the effect of each component of these mixed-model approaches in genetic association mapping. We generate simulated datasets based on HapMap data under various scenarios. Preliminary results indicate that the approaches based on mixed-model perform well in controlling for population stratification. However, these methods do not necessarily eliminate such confounding effect.