MDscan

MDscan identifies protein–DNA interaction sites at base-pair resolution by integrating ChIP-array data with sequence-motif discovery to map transcription factor binding.

Key Features:

• ChIP-Array Data Utilization: Leverages Chromatin Immunoprecipitation followed by cDNA microarray hybridization (ChIP-array) and ranking information to localize protein–DNA interactions and refine typical 1–2 kilobase resolution to the base-pair level.
• Motif Discovery: Combines word enumeration with position-specific weight matrix updating strategies, integrated with ChIP-array ranking to accelerate searches and improve motif-finding success rates.
• Accuracy and Validation: Demonstrated superior performance compared to several established motif-finding algorithms and recovered experimentally verified motifs in yeast examples such as STE12, GAL4, and RAP1 while predicting additional motifs including differential Rap1 binding at telomeres.
• Versatility: Applicable to ChIP-array experiments and other genomic studies where a subset of sequences is enriched for motif occurrences, enabling motif discovery across diverse datasets.
• Output: Reports putative motifs as position-specific probability matrices (position-specific weight matrices) and provides the individual sites used to construct these motifs along with their locations on input sequences.

Scientific Applications:

• Transcriptional regulation mapping: Identification of transcription factor binding sites to study gene expression regulation and transcriptional control.
• Comparative and differential binding analysis: Detection of differential protein–DNA binding across genomic regions, exemplified by Rap1 binding at telomeres.
• Regulatory motif discovery: Discovery of regulatory sequence motifs from ChIP-array or other experiments yielding enriched sequence subsets for downstream functional and comparative analyses.

Methodology:

Integrates ChIP-array ranking information with word enumeration and iterative position-specific weight matrix updates to identify motifs and output position-specific probability matrices and binding site locations.