Title: Towards site-based protein functional annotations

Authors: Seak Fei Lei, Jun Huan

Addresses: School of Electrical Engineering and Computer Science, University of Kansas, Lawrence, Kansas 66045, USA. ' School of Electrical Engineering and Computer Science, University of Kansas, Lawrence, Kansas 66045, USA

Abstract: The exact relationship between protein active centres and protein functions is unclear even after decades of intensive study. To improve functional prediction ability based on the local structures, we proposed three different methods. 1 We used Markov Random Field (MRF) to describe protein active region. 2 We developed filtering method that considers the local environment around the active sites. 3 We created multiple structure motifs by extending the motif to neighbouring residues. Our experiment results with enzyme families <40% sequence identity demonstrated that our methods reduced random matches and could improve up to 70% of the functional annotation ability (using area under curve).

Keywords: protein function prediction; MRF; Markov random field; protein functional annotations; protein active centres; protein functions; filtering; bioinformatics; multiple structure motifs; enzyme families; sequence identity; random matches.

DOI: 10.1504/IJDMB.2010.034200

International Journal of Data Mining and Bioinformatics, 2010 Vol.4 No.4, pp.452 - 470

Published online: 17 Jul 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article