BARTweb: a web server for transcription factor association analysis

Abstract

Identifying active transcription factors (TFs) that bind to cis-regulatory regions in the genome to regulate differential gene expression is a key task in gene regulation research. TF binding profiles from numerous existing ChIP-seq data can be utilized for association analysis with query data for TF identification, as alternative to DNA sequence motif analysis. Here, we present BARTweb, an interactive webserver for identifying TFs whose genomic binding patterns associate with input genomic features, by leveraging over 13,000 public ChIP-seq datasets for human and mouse. Using an updated Binding Analysis for Regulation of Transcription (BART) algorithm, BARTweb can identify functional TFs that regulate a gene set, or have a binding profile correlated with a ChIP-seq profile or enriched in a genomic region set, without a priori information of the cell type. Compared with the original BART package, BARTweb substantially reduces the execution time of a typical job by two orders of magnitude. We also show that BARTweb outperforms other existing tools in identifying true TFs from collected experimental data. BARTweb is a useful webserver for performing functional analysis of gene regulation. BARTweb is freely available at http://bartweb.org.