USP Electronic Research Repository

Advancing the accuracy of protein fold recognition by utilizing profiles from hidden Markov models

Lyons, J. and Dehzangi, A. and Heffernan, R. and Yang, Yuedong and Zhou, Y. and Sharma, Alokanand and Paliwal, K.K. (2015) Advancing the accuracy of protein fold recognition by utilizing profiles from hidden Markov models. IEEE Transactions on Nanobioscience, 14 (7). pp. 761-772. ISSN 1536-1241

[img] PDF - Accepted Version
Download (1MB)


Protein fold recognition is an important step towards solving protein function and tertiary structure prediction problems. Among a wide range of approaches proposed to solve this problem, pattern recognition based techniques have achieved the best results. The most effective pattern recognition-based techniques for solving this problem have been based on extracting evolutionary-based features. Most studies have relied on thePosition Specific Scoring Matrix (PSSM) to extract these features. However it is known that profile-profile sequence alignment techniques can identify more remote homologs than sequence-profile approaches like PSIBLAST. In this study we use a profile-profile sequence alignment technique, namely HHblits, to extract HMM profiles.We will show that unlike previous studies, using the HMM profile to extract evolutionary information can significantly enhance the protein fold prediction accuracy. We develop a new pattern recognition based system called HMMFold which extracts HMM based evolutionary information and captures remote homology information better than previous studies. Using HMMFold we achieve up to 93.8% and 86.0% prediction accuracies when the sequential similarity rates are less than 40% and 25%, respectively. These results are up to 10% better than previously reported results for this task. Our results show significant enhancement especially for benchmarks with sequential similarity as low as 25% which highlights the effectiveness of HMMFold to address this problem and its superiority over previously proposed approaches found in the literature.

Item Type: Journal Article
Additional Information: The HMMFold is available online at: =HMMFold.tar.bz2
Subjects: T Technology > T Technology (General)
Divisions: Faculty of Science, Technology and Environment (FSTE) > School of Engineering and Physics
Depositing User: Alokanand Sharma
Date Deposited: 06 Jan 2016 03:30
Last Modified: 12 Sep 2016 02:37

Actions (login required)

View Item View Item

Document Downloads

More statistics for this item...