Adaptive And Cooperative Harmony Search Models For RNA Secondary Structure Prediction

Loading...
Thumbnail Image
Date
2011-02
Authors
Mohammed A. Mohsen, Abdulqader
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Determining the function of RNA molecules relies heavily on its secondary structure. The current physical methods for secondary structure determination are expensive and time consuming. Several algorithms have been proposed for the RNA secondary structure prediction, including dynamic programming and metaheuristic algorithms. Harmony search (HS) is a new metaheuristic algorithm which succeeded in solving many different types of optimization problems. This research proposes three new variants of HS algorithm to address the RNA secondary structure prediction problem. The first variant is called HSRNAFold as a first application of HS for RNA secondary structure prediction. The second variant, AHSRNAFold, improves HSRNAFold by using adaptive parameter control. The third variant, CHSRNAFold, improves HSRNAFold by using a cooperative multiple harmony memories model. The behavior of the new HS variants is investigated and the impact of tuning the different parameters of these variants is evaluated. The experiments were conducted on 20 individuals with known structures from four RNA classes. The prediction accuracy was verified with native structures and other state-of-the-art algorithms. The results demonstrate that CHSRNAFold outperformed several state-of-the-art algorithms in terms of prediction accuracy.
Description
Keywords
Harmony Search Models
Citation