|
Activity Number:
|
464
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #308820 |
|
Title:
|
Gene Expression Data Analysis Using the Gene Ontology
|
|
Author(s):
|
Jiajun Liu*+ and Jacqueline M. Hughes-Oliver and J. Alan Menius
|
|
Companies:
|
North Carolina State University and North Carolina State University and GlaxoSmithKline
|
|
Address:
|
3009 Apt C Dorner Circle, Raleigh, NC, 27606,
|
|
Keywords:
|
Microarray ; Gene Ontology ; PLS ; Domain aggregation
|
|
Abstract:
|
New technologies for biological systems give scientists the ability to record thousands of measurements for each biomolecule including genes, proteins and metabolites. Domain enhanced analysis (DEA) uses the Gene Ontology to guide analysis of such data and increase interpretability. DEA uses a "top-down" approach to perform domain aggregation by first combining gene expressions related to each GO term using the Partial Least Squares (PLS) procedure. Then the first scores from the PLS procedure are used to test for differentially expressed patterns using the t test. We find the general t test inadequate for adjusting for the number of genes within each GO term. New tests are proposed by finding the true null distribution of each PLS score adjusted for the size of the GO term. We also discuss the impact of using different two-class classification response variables, namely 0/1 or -1/1.
|
