Joint Modeling of Epigenetic Data, Transcription Factor Binding Data, and Expression Data

• Activity Number: 189
• Type: Invited
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Guanghua Xiao*+
• Companies: UT Southwestern Medical Center
• Keywords: integrative analysis ; epigenetic data ; drug addiction ; bayesian modeling

Abstract

A large amount of genome-wide molecular profiling data has been accumulated on mRNA expression, histone modification, DNA methylation and transcription factor (TF) binding to study drug addiction. In this study, we develop a novel statistical model to joint model multiple molecular profiling data for detection of important genes associated with cocaine addiction. Here, data from ChIP-chip experiments include genome-wide DNA methylation, histone modification, TF binding data with cocaine exposure and with saline treatment. Expression data will include mRNA expression datasets with and without cocaine treatment. A Bayesian hierarchical model, combined with a logistic regression model for multiple-level responses, is used to integrate information from the multiple data sources available. Both simulation studies and real data examples show the good performance of the proposed model.