TFBSFootprinter

TFBSFootprinter predicts transcription factor binding sites in vertebrate species by scoring JASPAR matrices against Ensembl promoter sequences and integrating multiple contextual experimental datapoints into a single score per site.

Key Features:

• Multi-species analysis: Predicts transcription factor binding sites in up to 317 vertebrate species using Ensembl transcript IDs.
• JASPAR matrix scoring: Scores JASPAR position weight matrices against Ensembl promoter sequences to identify potential binding sites.
• Multiomic data integration: Combines seven contextual experimental datapoints including evolutionary conservation (GERP), CAGE-defined transcription start sites (FANTOM5), eQTLs (GTEx), ChIP-seq peaks (GTRD), ATAC-seq accessibility (ENCODE), DNase footprints, gene expression correlation (FANTOM5), and observed/expected CpG ratio (Ensembl) into a single score per site.
• Species-specific data availability: In humans, uses GERP, FANTOM5 transcription start sites, FANTOM5 expression correlation, GTRD ChIP-Seq metaclusters, ENCODE ATAC-Seq peaks, GTEx eQTLs, and Ensembl CpG ratio; in non-human vertebrates, uses GERP, proximity to transcription start sites, and CpG ratio.
• Optimized data combinations: Determines combinations of multiomic data that provide the best model of binding for each transcription factor.

Scientific Applications:

• Gene regulation research: Predicts functional transcription factor binding sites to understand eukaryotic gene regulatory mechanisms.
• Disease mechanism studies: Supports investigation of disease mechanisms through accurate prediction of transcription factor binding sites.
• Comparative genomics across vertebrates: Enables transcription factor binding site analysis in model organisms, domesticated animals, and wild animals for cross-species regulatory studies.
• Drug discovery: Facilitates drug discovery research by predicting transcription factor binding sites relevant to therapeutic targets.

Methodology:

TFBSFootprinter scores JASPAR matrices against Ensembl promoter sequences and combines the match with contextual experimental datapoints into a single score per site. Benchmarking was performed on a manually curated and experimentally verified dataset of transcription factor binding sites.