Statistical Models for Clustering of DNA Binding Motifs

• Activity Number: 198
• Type: Contributed
• Date/Time: Tuesday, August 10, 2004 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Shane Tyler Jensen*+ and Jun S. Liu
• Companies: Harvard University and Harvard University
• Keywords: DNA binding motifs ; Bayesian hierarchical modeling ; nonparametric clustering ; cross-species comparisons

Abstract

In order to elucidate detailed gene networks and mechanisms of cell regulation, biologists are interested in modeling the similarity between different motifs that represent the recognition sites of DNA binding proteins. The common structure between a set of motifs is investigated by a Bayesian nonparametric hierarchical model. This nonparametric model enables similar motifs to be partitioned together into clusters. This approach has substantial benefits over traditional clustering methods. The clustering inference gained can be built into more sophisticated search algorithms based on the posterior predictive distribution from our model in order to find binding motifs in large sequence datasets. Our strategy is also used to predict co-regulated genes on the basis of the clustering of motifs found in the regulatory region of genes that are conserved between different but closely related species.