Bicluster Coherency Measures for Gene Expression Data

Abstract

Many studies have been proposed to analyze gene expression microarray data, emphasizing on the identification of genes that show related functions over only subsets of different conditions. Detection of these homogenous genes is a crucial step in this analysis. One of the main approaches to achieve this task is biclustering, which is a time-consuming process that starts with the identifying sets of genes as seeds, expanding theses seeds using heuristic searches along with a measure of coherency to assess the quality of the resulting biclusters. The identification of the suitable coherency measure is a critical task, not only affecting the expansion of initial seed biclusters, but also the final shape of them. In this paper, a number of bicluster coherency measures for gene expression data are reviewed and analyzed from both analytical and mathematical aspects to help researchers in the choice of the right measure.