Hierarchical Clustering of DNA Microarray Data Using a Hybrid of Bacterial Foraging and Differential Evolution

Abstract

Microarray technology is one the most important advances in bioinformatics which allows the study of the expression levels of a large number of genes simultaneously. Data mining techniques have been widely applied in order to infer useful knowledge from DNA microarray data. One of these principle techniques is clustering which groups expressed genes according to their similarity. Hierarchical clustering is one of the main clustering methods which represents data in dendrograms. In a previous work the authors used the genetic algorithms to optimize the hierarchical clustering quality based on different clustering measures. In this paper we propose another optimization method based on a hybrid of differential evolution and bacterial foraging optimization algorithm to handle the optimization problem of hierarchical clustering of DNA microarray data. We show through experiments that this hybrid optimization method is more appropriate to tackle this problem than the one which uses the genetic algorithms, as this new method gives a better clustering quality according to different clustering measures.