Improving Co-expressed Gene Pattern Finding Using Gene Ontology

A semi-supervised gene co-expressed pattern finding method, PatGeneClus is presented in this paper. PatGeneClus attempts to find all possible biologically relevant gene coherent patterns from any microarray dataset by exploiting both gene expression similarity as well as GO-similarity. PatGeneClus uses a graph-based clustering algorithm called DClique to generate a set of clusters of high biological relevance. We establish the effectiveness of PatGeneClus over several benchmark datasets usingwell-known validity measures. The clusters obtained by PatGeneClus have been found to be biologically significant due to their high p-values, Q-values and clustalW scores.