An Immune System Inspired Algorithm for Protein Function Prediction

Abstract

An important problem in the field of bioinformatics research is assigning functions to proteins that have not been annotated. The extent, to which protein function is predicted accurately, depends largely on the Protein-Protein interaction network. It has been observed that bioinformatics applications are benefited by comparing proteins on the basis of biological role. Similarity based on Gene Ontology is a good way of exploring the above mentioned fact. In this paper we propose a novel approach for protein function prediction by utilizing the fact that most of the proteins which are connected in Protein-Protein Interaction network, tend to have similar functions. Our approach, an immune system-inspired meta-heuristic algorithm, known as Clonal Selection Algorithm (CSA), randomly associates functions to unannotated proteins and then optimizes the score function which incorporates the extent of similarity between the set of functions of unannotated protein and annotated protein. Experimental results reflect that our proposed method outperforms other state of the art algorithms in terms of precession, recall and F-value, when utilized to predict the protein function of Saccharomyces Cerevisiae.