Abstract 3214: BART: An integrative bioinformatics toolkit and web server for functional transcription factor prediction

Abstract

Abstract The goal of this bioinformatics tool is to predict functional transcription factors (TFs) from various queries of target genomic features. TFs play a key role in regulating gene expression in a dynamic chromatin structure. From a target gene set or a genomic region set, it is challenging to identify functional TFs that regulate these genes or bind in these regions, because of complexity of enhancer distributions and insufficient specificity of DNA sequence motifs. Here, we present BART (Binding Analysis for Regulation of Transcription), an interactive bioinformatics toolkit and web server to predict functional TFs that regulate a set of co-regulated genes or associate with a genomic profile. From a gene set, BART first uses a machine learning approach to generate a genome-wide cis-regulatory profile by leveraging hundreds of publicly available ChIP-seq profiles for active enhancer histone mark H3K27ac and DNase-seq profiles for genome-wide cis-regulatory elements. Then it identifies a list of TFs whose binding profiles most correlate with the cis-regulatory profile by comparing with tens of thousands of ChIP-seq datasets. BART can also identify TFs associated with differential chromatin interaction from a pair of three-dimensional (3D) genomic datasets, such as Hi-C. BART integrates over 12,000 ChIP-seq datasets for over 900 TFs for human and over 500 TFs for mouse to make accurate predictions. We demonstrate that BART outperforms existing methods in identifying the true TF from target gene sets. BART can be useful for any researcher in studying cancer gene regulation and cancer epigenetics, and demonstrates the power of utilizing public data for computational biology research. BART web server is freely available at: http://bartweb.org. Citation Format: Zhenjia Wang, Wenjing Ma, Yifan Zhang, Neal E. Magee, Yang Chen, Chongzhi Zang. BART: An integrative bioinformatics toolkit and web server for functional transcription factor prediction [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3214.