EXPLORING KEY GENE INTERACTIONS USING PARTICLE SWARM OPTIMIZATION

Abstract

Clustering is an exploratory method that is widely used for analyzing similarity of data objects. Clustering helps biologist in identifying functional similarity of genes. Most of the techniques employed for clustering genes need prior knowledge of the number of feasible clusters. Here we propose a novel hybrid approach towards gene clustering, which implements Particle Swarm Optimization (PSO) technique to find out closely related clusters by exploring the domain knowledge from gene ontology. The proposed approach is validated using the benchmark dataset and compared the performance with standard community detection algorithms. The results are promising and able to derive meaningful clusters from the dataset.