Rxivist logo

Discovery of Primary, Cofactor, and Novel Transcription Factor Binding Site Motifs by Recursive, Thresholded Entropy Minimization

By Ruipeng Lu, Eliseos Mucaki, Peter K Rogan

Posted 09 Mar 2016
bioRxiv DOI: 10.1101/042853 (published DOI: 10.1093/nar/gkw1036)

Data from ChIP-seq experiments can determine the genome-wide binding specificities of transcription factors (TFs) and other regulatory proteins. In the present study, we analyzed 745 ENCODE ChIP-seq peak datasets of 189 human TFs with a novel motif discovery method that is based on recursive, thresholded entropy minimization. This method is able to distinguish correct information models from noisy motifs, quantify the strengths of individual sites based on affinity, and detect adjacent cofactor binding sites that coordinate with primary TFs. We derived homogeneous and bipartite information models for 89 sequence-specific TFs, which enabled discovery of 24 cofactor motifs for 118 TFs, and revealed 6 high-confidence novel motifs. The reliability and accuracy of these models were determined via three independent quality control criteria, including the detection of experimentally proven binding sites, comparison with previously published motifs and statistical analyses. We also predict previously unreported TF cobinding interactions, and new components of known TF complexes. Because they are based on information theory, the derived models constitute a powerful tool for detecting and predicting the effects of variants in known binding sites, and predicting previously unrecognized binding sites and target genes.

Download data

  • Downloaded 697 times
  • Download rankings, all-time:
    • Site-wide: 32,880
    • In bioinformatics: 3,734
  • Year to date:
    • Site-wide: 64,080
  • Since beginning of last month:
    • Site-wide: 63,697

Altmetric data


Downloads over time

Distribution of downloads per paper, site-wide


PanLingua

Sign up for the Rxivist weekly newsletter! (Click here for more details.)


News