A Study of Biclustering Coherence Measures for Gene Expression Data

Biclustering algorithms have become one of the main tools for the analysis of gene expression data. They allow the identification of local patterns defined by subsets of genes and subsets of samples, which cannot be detected by traditional clustering algorithms. However, although useful, biclustering is a NP-hard problem. Therefore, the majority of biclustering algorithms look for biclusters optimizing a pre-established coherence measure. In the last 20 years, several heuristics and measures have been published for biclustering. However, most of these publications do not provide an extensive comparison of bicluster coherence measures on practical scenarios. To deal with this problem, this paper analyze the behavior of 15 bicluster coherence measures and external evaluation regarding 9 algorithms from the literature on gene expression datasets. According to the experimental results, there is no clear relation between these measures and assessment using information from gene ontology.