Computational analysis of biclustering algorithms for identification of co-expressed genes

Abstract

Biclustering is the popular data mining technique for the analysis of gene expression data. Recently, several biclustering algorithms have been published for finding the co-expressed genes. Some of the biclustering algorithms perform well on the particular aspect and not on all the aspects. So, to select the proper biclustering algorithm for gene expression data analysis becomes the challenging task. In this paper, computational analysis of six prominent biclustering algorithms has been done. The algorithms are Cheng and Church, Plaid Model, Bimax, FABIA, BICLIC, and ICS. Performance of the algorithms has been evaluated experimentally with the help of important performance measuring issues such as quality measures, bicluster type, bicluster size, noise, overlapping, output nature, biologically significant biclusters, comprehensive search, and biologically involved processes. From the computational analysis, important results have been pointed out which will be very helpful to the researcher for designing the robust biclustering algorithm or while selecting the biclustering algorithm for the analysis of gene expression data in biological applications.