|
Activity Number:
|
222
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #305121 |
|
Title:
|
Statistical Methods for Motif Detection Incorporating Structural Features of DNA
|
|
Author(s):
|
Mayetri Gupta*+
|
|
Companies:
|
The University of North Carolina at Chapel Hill
|
|
Address:
|
Department of Biostatistics, Chapel Hill, NC, 27599,
|
|
Keywords:
|
gene regulation ; motif discovery ; generalized hidden Markov models
|
|
Abstract:
|
Gene regulation is controlled by transcription factors (TFs) binding to specific sites (motifs) on the genome. Motif patterns tend to be short and degenerate, and similar patterns often occur outside protein-bound regions, leading genomic sequence-based motif prediction methods to have high false positive rates. Recently, new experimental methods studying genome-wide protein-DNA interactions suggest that packaging of chromatin is strongly associated with binding of TFs. We propose a Bayesian statistical model and a Monte Carlo method to infer nucleosome positioning from high density genome tiling arrays, and demonstrate that using the chromatin structure information significantly improves TF binding site predictions in an yeast genome example.
|
