|
Activity Number:
|
323
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #302178 |
|
Title:
|
A Markov Chain Monte Carlo Bayesian Network Approach To Infer Gene Co-Regulation Patterns
|
|
Author(s):
|
Younhee Ko*+ and Chengxiang Zhai and Sandra L. Rodriguez-Zas
|
|
Companies:
|
University of Illinois at Urbana-Champaign and University of Illinois at Urbana-Champaign and University of Illinois at Urbana-Champaign
|
|
Address:
|
1004 W Main St. #201, Urbana, IL, 61801,
|
|
Keywords:
|
Bayesian network ; Markov Chain Monte Carlo Gaussian mixture model ; gene expression ; gene network
|
|
Abstract:
|
Bayesian networks are a powerful framework to infer gene pathways from gene expression microarray experiments. A Markov Chain Monte Carlo (MCMC) implementation of Bayesian mixture networks was compared to an Expectation-Maximization (EM) implementation. Both implementations were applied to predict a signaling pathway using honey bee brain microarray data, and a cell communication pathway using mouse embryo microarray data. The MCMC estimates confirmed known gene relationships and uncover new relationships reported in the literature. The posterior probability distributions over candidate network structures had several local maxima. The lack of a single network structure clearly supported by the data may be due to the numerous conditions considered and limited information within condition. The posterior density estimates offered insights into the dynamic nature of gene networks.
|
