|
Activity Number:
|
372
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #304899 |
|
Title:
|
Bayesian Hierarchical Model of Gene Expression and Methylation Data Through EM Algorithm
|
|
Author(s):
|
Jaesik Jeong*+
|
|
Companies:
|
Indiana University Purdue University Indianapolis
|
|
Address:
|
, Indianapolis, IN, 46254,
|
|
Keywords:
|
Bayesian hierarchical model ; EM algorithm ; Microarray data
|
|
Abstract:
|
Over the past decade, microarray technologies have been improved. Several tens of thousands of genes are expressed by a state-of-art technology simultaneously. These gene expression data are used to find relation between disease and gene. If some relationship between gene and disease is found, it will be very helpful to development of medicine which can be used to treat such disease. Nowadays, the pattern of methylation draws attention to many researchers because it is believed that DNA methylation and gene expression have some causal relationship. We consider merging both gene expression data and DNA methylation and work on the modeling of hierarchical Bayesian model to find the causal relationship. We use EM algorithm to get parameter estimates in the model. We validate the appropriateness of our model through simulation study and apply our method to real microarray data.
|
