ModuleDigger

ModuleDigger detects statistically overrepresented combinations of transcription factor binding sites (TFBS), i.e., cis-regulatory modules (CRMs), in gene sets using itemset mining and a statistical scoring scheme to prioritize biologically relevant CRMs.

Key Features:

• Itemset mining strategy: Uses itemset mining to identify co-occurring TFBS patterns across gene sets.
• Statistical scoring scheme: Applies a statistical scoring scheme to evaluate overrepresentation and rank candidate CRMs.
• Computational efficiency: Improves computational efficiency relative to optimization-based methods, enabling analysis of large gene sets and many transcription factors.
• Support for extensive TF repertoires: Operates on TFBS for an extensive array of potential transcription factors.
• CRM prioritization: Prioritizes biologically valid CRMs within coregulated gene sets.
• Benchmarking on ChIP-Chip data: Evaluated on ChIP-Chip benchmark datasets and compared against other CRM detection methods to assess accuracy and scalability.

Scientific Applications:

• CRM discovery in eukaryotic regulation: Identification of cis-regulatory modules that mediate transcriptional responses in eukaryotes.
• Analysis of coregulated gene sets: Detection of overrepresented TFBS combinations to infer regulatory programs in coregulated genes.
• Large-scale TFBS analysis: Analysis of large TFBS datasets involving many transcription factors to characterize combinatorial regulation.
• Method benchmarking: Comparative evaluation of CRM detection approaches using ChIP-Chip-derived benchmark data.

Methodology:

Applies itemset mining to TFBS presence data, uses a statistical scoring scheme to assess overrepresentation of TFBS combinations, and evaluates performance on ChIP-Chip benchmark datasets against other CRM detection tools.