Discovering unknown human and mouse transcription factor binding sites and their characteristics from ChIP-seq data

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Transcription factor binding sites (TFBSs) are crucial for gene regulation but are not fully characterized.
• Existing methods for identifying TFBSs from ChIP-seq data have limitations in quality control and discovery.

Purpose of the Study:

• To develop and validate a computational pipeline for discovering and characterizing novel TFBSs from ChIP-seq data.
• To re-evaluate quality control metrics for ChIP-seq experiments and improve TFBS discovery.

Main Methods:

• Developed a computational pipeline for analyzing ChIP-seq data to identify TFBS motifs.
• Applied the pipeline to human and mouse ENCODE ChIP-seq datasets.
• Utilized motif occurrences in peak regions as a quality control criterion, outperforming irreproducible discovery rate.

Main Results:

• Identified 2,058 motifs for human TFs and 163 for mouse TFs, with 487 novel motifs discovered.
• Found high TFBS density near transcription start sites, but most TFBSs (70%) are in introns and intergenic regions.
• Inferred frequent TF interactions, including 142 cobinding and 186 tethered binding TF pairs.

Conclusions:

• The developed pipeline significantly expands the repertoire of known TFBS motifs.
• TFBS distribution is broader than previously thought, with substantial numbers in intronic and intergenic regions, including enhancers.
• The findings provide new insights into TF binding, interactions, and genome-wide TFBS organization.