Improved Protein Structural Class Prediction Using Genetic Algorithm and Artificial Immune System

Abstract

Predicting the structure of a protein from primary sequenceis one of the challenging problems in Molecularbiology. In this context, protein structural class informationprovides a key idea of their structure and also other featuresrelated to the biological function. In this paper we present anew optimization approach based on Genetic algorithm (GA)and artificial immune system (AIS) for predicting the proteinstructural class. It uses the maximum component coefficientprinciple in association with the amino acid compositionfeature vector to efficiently classify the protein structures.The effectiveness is evaluated by comparing the results withthat obtained from other existing methods using a standarddatabase. Especially for all and + class protein, therate of accurate prediction by the proposed methods is muchhigher than their counterparts.