Detecting Genetic Association in Case-control Studies Using Similarity-based Association Tests

• Activity Number: 265
• Type: Invited
• Date/Time: Wednesday, August 14, 2002 : 8:30 AM to 10:20 AM
• Sponsor: International Chinese Statisticial Association
• Author(s): Shuanglin Zhang*+ and Kenneth Kidd and Hongyu Zhao
• Affiliation(s): Michigan Technological University and Yale University and Yale University
• Keywords: case-control studies ; population stratificatio ; coalescent models ; population genetics

Abstract

Although traditional case-control studies may be subject to bias caused by population stratification, alternative methods such as family-based methods may be less powerful due to overmatching between cases and controls, and may not be easy for sample collection. Recently, several statistical methods have been proposed for association tests in structured populations using case-control designs that may be robust to population stratification. In this article, we propose a similarity-based association test (SAT) to identify association between a candidate marker and a disease of interest using case-control designs. We first determine whether two individuals are from the same subpopulation or not using genotype data at a set of independent markers. We then perform an association test by comparing within-subpopulation allele-frequency differences between cases and controls. Simulation results show that the SAT has correct type-I error rate in the presence of population stratification. The power of the SAT is higher than that using family-based association designs and is also higher than other robust association methods when the high-risk allele is the same across all subpopulations.