A Bayesian Approach to Informatively Clustering Microarray Data

• Activity Number: 657
• Type: Contributed
• Date/Time: Thursday, August 4, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Jessie Hsu*+ and David Alan Schoenfeld and Dianne M. Finkelstein
• Companies: Harvard University and Massachusetts General Hospital and Massachusetts General Hospital/Harvard School of Public Health
• Keywords: Microarray ; Bayesian Inference ; Clustering

Abstract

Outcome-driven clustering of gene expression data has the potential to produce clusters that are more useful for describing the clinical course of disease. We propose a joint hierarchical model that assumes a relationship between gene clusters and a continuous patient outcome. Gene expression is modeled using normally distributed latent cluster random effects such that genes in the same cluster are correlated, and a linear combination of these random effects is used to describe clinical outcome. To fit the model, we implement a Markov chain Monte Carlo algorithm based on a likelihood where the random effects are integrated out. The advantage of the Bayesian approach is that it simultaneously finds the gene clusters and estimates the model parameters, allowing the clusters to be informed by patient outcome and subsequently resulting in better predictors of outcome. We evaluate the effect of including outcome in the model through simulation studies, as well as describe a strategy for prediction. These methods are applied to trauma data from the Inflammation and Host Response to Injury research program.