|
Activity Number:
|
421
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #309448 |
|
Title:
|
A Bayesian Approach to Pool Multiple Datasets for Periodically Expressed Genes Detection
|
|
Author(s):
|
Xiaodan Fan*+ and Jun S. Liu
|
|
Companies:
|
Harvard University and Harvard University
|
|
Address:
|
1 Oxford Street, Cambridge, MA, 02138,
|
|
Keywords:
|
reversible jump MCMC ; Bayesian meta-analysis ; cell cycle gene ; transformation group ; time series ; model selection
|
|
Abstract:
|
Multiple microarray time series datasets have been produced in an effort to detect periodically expressed genes. There are discrepancies with regard to the identity and numbers of periodically expressed genes. We proposed a hierarchical model to pool all datasets to get better classification of genes as either periodically expressed (PE) or aperiodically expressed (APE). A damping sinusoidal function with linear tread and iid Gaussian noise is used to model each PE time series. APE time series are simple modeled by a linear tread plus iid Gaussian noise. Within an experiment, all genes share the same period. Between experiments, all genes share the same phase shift. MCMC is used to estimate the parameters. Reversible Jump MCMC is used to dynamically select models (PE or APE) for each time course. Transformation move, group move, block swapping are used to improve the mixing of the chain.
|
