The p53HMM algorithm: using profile hidden markov models to detect p53-responsive genes

Todd Riley,Xin Yu,Eduardo D Sontag ,Arnold J Levine

doi:10.1186/1471-2105-10-111

Abstract

BackgroundA computational method (called p53HMM) is presented that utilizes Profile Hidden Markov Models (PHMMs) to estimate the relative binding affinities of putative p53 response elements (REs), both p53 single-sites and cluster-sites. These models incorporate a novel "Corresponded Baum-Welch" training algorithm that provides increased predictive power by exploiting the redundancy of information found in the repeated, palindromic p53-binding motif. The predictive accuracy of these new models are compared against other predictive models, including position specific score matrices (PSSMs, or weight matrices). We also present a new dynamic acceptance threshold, dependent upon a putative binding site's distance from the Transcription Start Site (TSS) and its estimated binding affinity. This new criteria for classifying putative p53-binding sites increases predictive accuracy by reducing the false positive rate.ResultsTraining a Profile Hidden Markov Model with corresponding positions matching a combined-palindromic p53-binding motif creates the best p53-RE predictive model. The p53HMM algorithm is available on-line: ConclusionUsing Profile Hidden Markov Models with training methods that exploit the redundant information of the homotetramer p53 binding site provides better predictive models than weight matrices (PSSMs). These methods may also boost performance when applied to other transcription factor binding sites.

Highlights

A computational method is presented that utilizes Profile Hidden Markov Models (PHMMs) to estimate the relative binding affinities of putative p53 response elements (REs), both p53 single-sites and cluster-sites
Since insertions and deletions throw off the reading frame of a weight matrix, any PSSM approach will inherently mis-score at least 35% of these 20 sites
A novel training method that boosts predictive power To increase the predictive power of our p53-motif PHMMs, we attempt to exploit the a priori knowledge that when proteins bind as homodimers or homotetramers, their corresponding binding sites typically have a palindromic, repeat, and/or reverse complement structure

Summary

Introduction

A computational method (called p53HMM) is presented that utilizes Profile Hidden Markov Models (PHMMs) to estimate the relative binding affinities of putative p53 response elements (REs), both p53 single-sites and cluster-sites. These models incorporate a novel "Corresponded Baum-Welch" training algorithm that provides increased predictive power by exploiting the redundancy of information found in the repeated, palindromic p53-binding motif. Alignments of the 160 experimentally validated p53 binding sites reveal that any PSSM approach would inherently mis-score at least 30% of them as well Another observation is that additional p53 half-sites are immediately adjacent (in yellow) to the ones used to define the consensus in 15 of the 20 target sites (75%). Since the genome-wide immunoprecipitation study was designed to pull down the highest affinity sites, the fact that 75% of the target sites are p53 cluster-sites is the first indication that cluster-sites of 3 or more half-sites confer higher binding affinity [22]

Methods

Results

Conclusion