Bayesian approach for detecting differentials of gene expression with the mixture prior

Abstract

In DNA microarray analysis, it is an important problem to detect differentials of gene expression. We use the gamma and the weibull distributions in modeling gene expression. We assume mixture priors on the parameters representing different effects between two experimental conditions. Markov chain Monte Carlo methods are used to compute the Bayes factor and posterior means. We perform a simulation study and real data analysis to demonstrate our theoretical results.