|
Activity Number:
|
296
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #302823 |
|
Title:
|
Methods for Identifying the Genetic Variants Associated with High-Order Expression Modules
|
|
Author(s):
|
Hongzhe Li*+
|
|
Companies:
|
University of Pennsylvania
|
|
Address:
|
Department of Biostatistics, Philadelphia, PA, 19104,
|
|
Keywords:
|
eQTL ; Regularization ; Variable Selection ; Regression on Correlation
|
|
Abstract:
|
Many recent studies have shown that the level of expression of gene can be heritable and many eQTLs have been identified by traditional linkage/association analysis. These studies mainly focus on the effects of genetic variants on the levels of expression of genes. However, genetic variants can also affect expression coordination among genes and therefore high-order expression modules. This paper proposes a regression-on-correlation (RegCor) model to link the genetic variants to expression association between two genes in order to identify the variants that are association with high-order expression modules. A score test and a regularized MLE procedure are proposed for hypothesis testing and for selection of genetic variants. Some theoretical results, simulations and application to expression data in yeast segregants will be presented.
|
