|
Activity Number:
|
225
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #305576 |
|
Title:
|
A Bayesian Semiparametric Hierarchical Model for Analyzing Differential Expression in Sequence-Based Gene Expression Data
|
|
Author(s):
|
Soma S. Dhavala*+ and Bani K. Mallick
|
|
Companies:
|
Texas A&M University and Texas A&M University
|
|
Address:
|
1100 Hensel Dr., College Station, TX, 77840,
|
|
Keywords:
|
Dirichlet Process Prior ; Differential Gene Expression ; Markov Chain Monte Carlo ; Bayesian hierarchical models ; Semiparametric Modeling
|
|
Abstract:
|
In this paper, we propose a Bayesian Semiparametric Hierarchical model for analyzing differential expression (DE) in sequence-based gene expression data. The count data is modeled using a zero-inflated Poisson likelihood. Dependency among the genes is modeled by eliciting a Dirichlet process prior thru this DPP. A finite normal mixture base distribution is elicited to facilitate the computation of DE probabilities and control the false discovery rate. The posterior inference is carried-out using Markov Chain Monte Carlo simulations. We applied the model to analyze Salmonella infection gene expression data in Bovines produced massively parallel signature sequencing technology. Differentially expressed genes obtained with our methodology are validated against the existing literature. The clustering information is useful in unraveling the functionality of some undocumented signatures.
|
