|
Activity Number:
|
386
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #306699 |
|
Title:
|
Bayesian Clustering of Short Temporal Gene Expression Dynamics
|
|
Author(s):
|
Ling Wang*+ and Paola Sebastiani and Marco Ramoni
|
|
Companies:
|
Boston University and Boston University and Harvard Medical School
|
|
Address:
|
401 Broadway, Cambridge, MA, 02139,
|
|
Keywords:
|
Bayesian model ; invariance ; polynomial model ; caged
|
|
Abstract:
|
We present an extension of CAGED (Clustering Analysis of Gene Expression Dynamics) to cluster gene expression profiles measured in short temporal/dose-response microarray experiments. In contrast to the initial version of CAGED, in which temporal expression profiles are modeled by autoregressive equations, our new algorithm describes the trend using polynomial models of time/dosage. Our Bayesian approach uses proper conjugate priors for the model parameters so that the algorithm is invariant to linear reparameterizations of time/dosage. We compare our approach with the recently introduced program STEM (Short Time-series Expression Miner) to show that our method can find the correct number of clusters and allocate gene expression profiles to the right clusters in simulation studies, and produce more biologically meaningful Gene Ontology enriched clusters in real dataset.
|
