RPMCMC

RPMCMC detects sequence motifs in large nucleotide datasets by running a repulsive parallel Markov chain Monte Carlo approach to improve motif diversity and detection accuracy.

Key Features:

• Parallel Processing: Implements a parallel version of the Gibbs motif sampler with multiple interacting motif samplers running concurrently.
• Repulsive Force Mechanism: Applies a repulsive force among different motif samplers to discourage convergence to similar motifs and promote exploration of distinct motif solutions.
• Enhanced Detection Accuracy: Emphasizes detection precision to identify motifs that are often missed by conventional methods.
• Application to Large Datasets: Designed to handle large-scale genome-wide ChIP-seq datasets such as those from the ENCODE project.

Scientific Applications:

• ENCODE ChIP-seq analysis: Applied to 228 transcription factor ChIP-seq datasets from the ENCODE project for motif discovery.
• Cofactor motif discovery: Detects reliable cofactor interacting motifs in genome-wide ChIP-seq data that were previously undetectable by conventional methods.

Methodology:

Runs parallel interacting Gibbs motif samplers within a Markov chain Monte Carlo framework and uses a repulsive force among samplers to encourage exploration of distinct regions of motif space.