Identification of Biologically Relevant Biclusters from Gene Expression Dataset of Duchenne Muscular Dystrophy (DMD) Disease Using Elephant Swarm Water Search Algorithm

Abstract

A gene expression dataset contains expressions of different genes at different conditions of a disease. The biclustering procedure explores only those subsets of genes, which have similarities in expression behaviors across some subsets of conditions. These subsets of genes and conditions form sub-matrices, known as biclusters. Here, for the first time, the biclustering approach is applied to Duchenne Muscular Dystrophy (DMD) disease dataset. This paper presents a meta-heuristic-based method for identifying biclusters from the dataset related to Duchenne Muscular Dystrophy (DMD) disease. It identifies shifting and scaling pattern-based biclusters considering different objectives together. For this purpose, Elephant Swarm Water Search Algorithm (ESWSA) and a proposed variant of ESWSA, named Chaotic Move Elephant Swarm Water Search Algorithm (CM-ESWSA) have been implemented. The proposed method (CM-ESWSA) has been able to recognize the shifting and scaling pattern-based biclusters of better quality. To determine the efficiency of ESWSA and CM-ESWSA, statistical testing and benchmark analysis have also been done and it shows that the proposed method outperforms the basic ESWSA.