|
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 - #309149 |
|
Title:
|
Bayesian Pathway Annotation Analysis of Genomewide Expression Profiles
|
|
Author(s):
|
Haige Shen*+ and Mike West
|
|
Companies:
|
Duke University and Duke University
|
|
Address:
|
214A Old Chemistry Bldg, Durham, NC, 27708-0251,
|
|
Keywords:
|
gene set enrichment analysis ; marginal likelihood ; pathway annotation ; variational approximations
|
|
Abstract:
|
Understanding molecular pathways underlying cancer phenotypes is essential to uncovering dynamic process of tumorigenesis. As part of this, linking quantified, experimentally defined gene expression signatures with known biological pathway gene sets is a key challenge. We develop a Bayesian approach to the problem of genome-wide expression-based pathway annotation. This involves a model-based approach to matching experimental signatures of structure or outcomes in gene expression---represented in terms of ranked and weighted gene lists---to multiple pathway gene sets from curated databases. One overall result is posterior probabilities over pathways for each experimental signature. We discuss the modeling approach and some computational challenges, and demonstrate the use of MCMC and variational methods that provide solution. Examples in cancer pathway analysis highlight the approach.
|
