|
Activity Number:
|
403
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #301916 |
|
Title:
|
Surrogate Variable Analysis
|
|
Author(s):
|
Jeffrey Leek*+ and John Storey
|
|
Companies:
|
Mount Sinai School of Medicine and Princeton University
|
|
Address:
|
One Gustave L. Levy Place, New York, NY, 10029,
|
|
Keywords:
|
multiple testing dependence ; surrogate variables ; gene expression ; principal component analysis
|
|
Abstract:
|
High-throughput experiments measure data for thousands of related features and seek to rank them for association with variables of experimental or clinical importance. The process of ranking features for association with primary variables is complicated by other factors that may influence thousands of features at a time. In high-dimensional experiments these factors are often unknown, unmeasured, or incapable of being tractably modeled. We provide a statistical framework for modeling large-scale noise dependence caused by unmodeled factors in high-throughput data. We introduce surrogate variables, estimable linear combinations of the true unmodeled factors, that can be included when modeling the relationship between the primary variables and the feature data. We propose algorithms for estimating surrogate variables based on principal component analysis of subsets of features.
|
