Semiparametric Functional Mapping for Irregular Longitudinal Data Using Penalized Spline and MCMC

• Activity Number: 530
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Kiranmoy Das*+
• Companies: Penn State
• Keywords: Semiparametric ; Penalized spline ; MCMC ; Bayes factor ; QTL ; Functional mapping

Abstract

Modelling the covariance structure for irregular sparse longitudinal data has been a challenging issue in recent days. Nonnegative definiteness property of the covariance matrix makes it difficult to model for high dimensional data. We present a simple semiparametric approach to model mean-covariance structure jointly for such data. Penalized spline is used to model the mean function while an extended generalised linear model is used to model covariance matrix. parameters are estimated by MCMC, using Gibb's sampling as well as Metropolis Hastings algorithm. We analyzed BMI data to test if there is a genetic factor which controls BMI at different ages. In the functional mapping framework, we derive the full conditional distributions for the mean and covariance parameters and compute Bayes factor to test our hypothesis.