A Novel Bayesian Approach for Differential Expression Analysis with RNA-Seq Data

• Activity Number: 527
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Peng Liu*+ and Fangfang Liu and Chong Wang
• Companies: Iowa State University and Iowa State University and Iowa State University
• Keywords: Bayesian ; differential expression ; Dirichlet process mixture model ; optimal test ; RNA-seq

Abstract

RNA-sequencing (RNA-seq) technologies have revolutionized the way biologists study gene expression and generated tremendous amount of data waiting for analysis. Detecting differentially expressed genes is one of the fundamental steps in RNA-seq data analysis. We model the count expression data for each gene using a Poisson-Gamma hierarchical model, or equivalently, a negative binomial model. In this paper, we propose an optimal test using Dirichlet process mixture models for the parameter corresponding to the fold change between the two treatment means and conjugate priors for the other parameters in our model. We develop an inference strategy using collapsed Gibbs algorithm for differential expression analysis. Simulation results show our method is promising.