Finding Sequence Motifs with Bayesian Models Incorporating Positional Information

• Activity Number: 28
• Type: Contributed
• Date/Time: Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
• Sponsor: ENAR
• Author(s): Nak-Kyeong Kim*+ and John L. Spouge
• Companies: Old Dominion University and The National Center for Biotechnology Information
• Keywords: Bioinformatics ; Computational biology ; Bayesian model ; Motif identification ; Gibbs sampling ; Robustness

Abstract

Biologically active sequence motifs often have positional preferences with respect to a genomic landmark. For example, many known transcription factor binding sites (TFBSs) occur within several hundred bases upstream of a transcription start site (TSS). In this study, we propose Bayesian models that systematically combine sequence and positional information. Predictions with and without positional information were compared on two human TFBS data sets, each containing sequences corresponding to the interval [-2000, 0] bases upstream of a known TSS. A rigorous statistical analysis showed that positional information significantly improved the prediction of sequence motifs, and an extensive cross-validation study showed that the proposed model was robust against mild misspecification of its parameters.