Pathway Marker Identification Using Gene Expression Data Analysis: A Particle Swarm Optimisation Approach

A pathway is a set of genes together performing similar biological functions. All genes in a pathway may not participate in a cellular function but some of them have strong association. The aim of this paper is to identify those genes having high degree of association in a pathway. Here, Mutated Binary Particle Swarm Optimization (MBPSO) has been proposed and implemented to identify differentially expressed genes in a pathway. The active genes of this selected pathway are used for sample classification using Support Vector Machine (SVM). 10-fold cross validation has been applied here to compute different performance metrics to assess the quality of the classifier. The proposed MBPSO has been applied to three microarray gene expression datasets to evaluate its performance with respect to the state-of-the-art algorithms. It is found that MBPSO outperforms other state-of-the-art algorithms in most of the performance metrics. Disease-gene association is examined through DisGeNET software. It is found that selected genes are biologically relevant. Some of the most relevant genes are CCND1, CAV1, IGF1, PIK3CB, MDM2, MMP9, VEGFA.