Mining positive and negative co-regulation patterns from microarray data

Currently, pattern-based and tendency-based models are very popular for clustering co-regulated genes. In this paper we propose another novel model, namely g-Cluster. The proposed model has the following advantages: (1) find positive and negative co-regulated genes in a shot, (2) get away from the restriction of magnitude transformation relationship among genes, and (3) guarantee quality of clusters and significance of regulations using a novel similarity measurement gCode and two user-specified thresholds, called wave constraint threshold and regulation threshold respectively. We also design a novel tree-based clustering algorithm, FBTD, combined with efficient pruning rules to identify all maximal g-Clusters. The extensive experiments on real and synthetic datasets show that (1) our algorithm can effectively and efficiently find an amount of co-regulated gene clusters missed by previous models, which are potentially of high biological significance, and (2) our algorithm is superior to the existing approaches.