Empirical Likelihood in Genetic Mixture Model

• Activity Number: 380
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: IMS
• Author(s): Jing Qin*+ and Pengfei Li and Yukun Liu
• Companies: NIH and University of Waterloo and East China Normal University
• Keywords: Empirical likelihood ; mixture model

Abstract

In genetic experiments, data are often collected from complex mixtures of distribution functions with known mixing proportions. The well-known examples include plants and animal quantitative trait locus(QTL) studies and human genetic linkage analysis. In this talk we explore a semiparametric model used by Anderson (1979), in which the log ratio of probability (density) functions from components is linear in the observations. Since the Fisher information is degenerated under the null that there is no mixture, we have to expand the likelihood ratio up to 4-th order for the limiting distribution Simulation results and a real cancer study application data are presented.