Prediction of transcription factor families using DNA sequence features

Ashish Anand*, Gary B. Fogel, Pugalenthi Ganesan, P. N. Suganthan

*Corresponding author for this work

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

    1 Scopus citations

    Abstract

    Understanding the mechanisms of protein-DNA interaction is of critical importance in biology. Transcription factor (TF) binding to a specific DNA sequence depends on at least two factors: A protein-level DNA-binding domain and a nucleotide-level specific sequence serving as a TF binding site. TFs have been classified into families based on these factors. TFs within each family bind to specific nucleotide sequences in a very similar fashion. Identification of the TF family that might bind at a particular nucleotide sequence requires a machine learning approach. Here we considered two sets of features based on DNA sequences and their physicochemical properties and applied a one-versus-all SVM (OVA-SVM) with class-wise optimized features to identify TF family-specific features in DNA sequences. Using this approach, a mean prediction accuracy of ~80% was achieved, which represents an improvement of ~7% over previous approaches on the same data.

    Original languageEnglish (US)
    Title of host publication3rd IAPR International Conference on Pattern Recognition in Bioinformatics, PRIB 2008
    Pages154-164
    Number of pages11
    DOIs
    StatePublished - Dec 8 2008
    Event3rd IAPR International Conference on Pattern Recognition in Bioinformatics, PRIB 2008 - Melbourne, VIC, Australia
    Duration: Oct 15 2008Oct 17 2008

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume5265 LNBI
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349

    Other

    Other3rd IAPR International Conference on Pattern Recognition in Bioinformatics, PRIB 2008
    CountryAustralia
    CityMelbourne, VIC
    Period10/15/0810/17/08

    Keywords

    • Multi-class classification
    • Transcription factor family prediction

    ASJC Scopus subject areas

    • Theoretical Computer Science
    • Computer Science(all)

    Fingerprint Dive into the research topics of 'Prediction of transcription factor families using DNA sequence features'. Together they form a unique fingerprint.

    Cite this