|
Activity Number:
|
418
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #305964 |
|
Title:
|
Using Microarray Gene-Coexpression Networks To Increase Gene Screening Validation Success and To Build Accurate Classifiers
|
|
Author(s):
|
Wei Zhao*+ and Steve Horvath and Paul Mischel and Aldons J. Lusis and Stanley Nelson
|
|
Companies:
|
University of California, Los Angeles and University of California, Los Angeles and University of California, Los Angeles and University of California, Los Angeles and University of California, Los Angeles
|
|
Address:
|
Department of Human Genetics, Los Angeles, CA, 90095,
|
|
Keywords:
|
cancer ; prediction ; microarray ; co-expression network ; screening ; validation
|
|
Abstract:
|
Microarray gene expression profiles have begun to prove useful for classifying subsets of patients or tumors, and for predicting survival and response to therapy. However, genes identified as predictive in one microarray study have frequently failed to validate in other studies. Traditional gene screening methods often select genes by correlating the expression profiles with microarray sample trait information (e.g. patients survival or case-control status). Several groups demonstrated that selecting genes simply on the basis of a p-value or fold-change criterion may lead to a gene list with poor validation success Michiels et al. (2005). Further, standard classifiers (e.g. k-nearest neighbor) may have poor accuracy in independent data. We propose to use a gene network-based gene screening strategy for identifying gene candidates. We provide ample empirical evidence of the usefulness.
|
