BART

BART predicts transcription factors that regulate a given gene set or associate with a genomic/epigenomic profile by leveraging large-scale ChIP-seq data for functional binding inference.

Key Features:

• Input types: Accepts gene sets or genomic/epigenomic profiles as queries for transcription factor prediction.
• ChIP-seq compendium: Leverages a repository of over 6,000 ChIP-seq datasets covering over 400 transcription factors in human and mouse.
• Correlation-based inference: Ranks candidate transcription factors by correlating their ChIP-seq binding profiles with the input gene sets or genomic profiles.
• Distal enhancer detection: Infers functional binding at distal enhancers, addressing limitations of DNA sequence motif analysis.
• Performance: Aims for high sensitivity and specificity in transcription factor prediction by exploiting large-scale empirical binding data.
• Implementation: Implemented in Python and built to operate on publicly available ChIP-seq datasets.

Scientific Applications:

• Regulatory mechanism discovery: Identifies transcription factors underlying gene regulatory programs in development.
• Disease research: Predicts transcription factors associated with gene expression changes in disease progression.
• Cell differentiation studies: Infers transcription factors driving cellular differentiation processes.
• Hypothesis generation: Produces candidate transcription factor–gene interactions for experimental validation.

Methodology:

Correlates input gene sets or epigenomic profiles with a compendium of over 6,000 public ChIP-seq datasets for over 400 transcription factors in human and mouse; implemented in Python.