Hierarchical State-Space Model for Microarray Short Time Course Experiments

• Activity Number: 56
• Type: Topic Contributed
• Date/Time: Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Haiyan Wu*+ and Ming Yuan and Susan Kaech and M. Elizabeth Halloran
• Companies: Emory University and Georgia Institute of Technology and Yale University and Fred Hutchinson Cancer Research Center
• Keywords: microarray ; time course ; hidden Markov model ; empirical Bayes ; informative prior

Abstract

Microarray time course experiments are important in investigating dynamic biological processes. Four common tasks in microarray time course experiments are hypothesis testing, pattern recognition, trajectory formation, and network/pathway reconstruction. Naïve approaches, treating time series as independent, tend to ignore the correlation information across time and suffer from lower sensitivity and higher misclassification rate. Model-based approaches have been proposed to rank/identify temporally differentially expressed genes. However, most of the up-to-date models are inappropriate for temporal microarray experiments with short time series. Here, we introduce a hierarchical model integrated with Hidden Markov Model structure and auto-regression to model the temporal gene expression profile on both expression and state levels to achieve these four common tasks.