Analysis of Genomic Data via Likelihood Ratio Test in Composite Kernel Machine Regression

• Activity Number: 420
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: ENAR
• Author(s): Ni Zhao* and Michael C. Wu
• Companies: Fred Hutchinson Cancer Research Center and Fred Hutchinson Cancer Research Center
• Keywords: likelihood ratio test ; kernel machine regression ; multiple kernels

Abstract

Semiparametric kernel machine regression has emerged as a powerful and flexible tool in genomic studies in which genetic variants are grouped into biologically meaningful entities for association testing. Recent advances have expanded the method to test for the effect of multiple groups of genomic features via a composite kernel that is constructed as a weighted average of multiple kernels. Variance component testing is used to evaluate the significance but requires fixing the weighting parameters or perturbation. In this paper, we focus on the (restricted) likelihood ratio test for kernel machine regression with composite kernels where instead of fixing the weighting parameter, we estimate the weighting parameter by maximizing the likelihood functions through the linear mixed model with multiple variance components. We derive the spectral representation of (R)LRT in linear mixed models with multiple variance components to obtain their finite sample distribution. We conduct extensive simulations to evaluate the power and type I error. Finally, we applied to proposed (R)LRT method to a real study to illustrate our methodology