Gaussian Process--Based Semiparametric Bayesian QTL Mapping for Longitudinal Data

• Activity Number: 652
• Type: Contributed
• Date/Time: Thursday, August 8, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Wonil Chung*+ and Fei Zou
• Companies: The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill
• Keywords: Gaussian process ; Bayesian ; mixed model ; longitudinal Data ; QTL ; Cholesky decomposition

Abstract

In this paper, we extend the Gaussian process based nonparametric Bayesian variable selection method to handle longitudinal Data. To further improve the flexibility of the Bayesian model, we propose a grid-based method to model the covariance structure of the data, which can accurately approximate any complex covariate structure.The dimension of the working covariance matrix depends on the number of fixed grid points even though each subject may have different number of measurements at different time points. We apply Chen and Dunson's method with a modified Cholesky decomposition to estimate the covariance of random effect. To draw MCMC samples of parameters, both hybrid Monte Carlo method and Gibbs sampling method are used. The performance of the proposed method is evaluated by simulations. We analyze real mouse data on age-related body weight in backcross mice with all time points together. Our semiparametric Bayesian analysis identified strong evidence of QTL activity on chromosome 1.