SeqSite

SeqSite detects and localizes transcription factor binding sites (TFBSs) from ChIP-seq data using a two-step, strand-specific modeling strategy to improve precision of individual binding site localization.

Key Features:

• Two-step strategy: Combines detection of tag-enriched regions with strand-specific density modeling to pinpoint TFBSs.
• Tag enrichment detection: Identifies potential binding regions by clustering ChIP-seq tags and applying statistical hypothesis testing.
• Density-profile modeling: Models the density profile of tags and refines site locations using least-squares model fitting on each DNA strand.
• Strand-specific merging: Merges detections from both strands to determine individual binding site positions.
• Resolution of adjacent sites: Accurately locates closely spaced adjacent binding sites, addressing overlapping ChIP-seq signals.
• Simulated-data validation: Demonstrated ability to identify binding sites separated by more than 40 base pairs in simulated datasets.
• Application to human datasets: Applied to three human transcription factor ChIP-seq datasets for comparative analysis of localization resolution.

Scientific Applications:

• High-resolution TFBS mapping: Precise localization of TFBSs from ChIP-seq data for genome-wide binding site annotation.
• Analysis of closely spaced binding events: Discrimination of adjacent binding sites to study cooperative binding and fine-scale regulatory architecture.
• Method benchmarking: Comparative evaluation of TFBS detection performance using simulated data and human transcription factor ChIP-seq datasets.

Methodology:

Detect tag-enriched regions by clustering ChIP-seq tags and applying statistical hypothesis testing; model tag density profiles and fit least-squares models on each DNA strand; merge strand-specific detections to pinpoint individual binding sites.