|
Activity Number:
|
313
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #308014 |
|
Title:
|
Associating High-Dimensional Response and Covariate Data
|
|
Author(s):
|
Mahlet G. Tadesse*+ and Stefano Monni
|
|
Companies:
|
University of Pennsylvania and University of Pennsylvania
|
|
Address:
|
Department of Biostatistics & Epidemiology, Philadelphia, PA, 19104-6021,
|
|
Keywords:
|
Markov chain Monte Carlo ; high-dimensional data
|
|
Abstract:
|
In recent years, high-dimensional datasets have become common in many applications. Several methods have been developed to exploit the wealth of information in these data. Most of the procedures have focused on relating the high-dimensional covariates to univariate outcomes, or at most a few correlated outcomes. Often, however, both the response and covariate data are high-dimensional. We propose a method to identify subsets of covariates associated with sets of correlated outcomes. We illustrate the method with an application to eQTL analysis, in which gene expression microarray data are related to genotype data from thousands of SNP markers. The method allows the identification of correlated gene expression patterns modulated by sets of DNA variations.
|
