DCoSpect: A Novel Differentially Coexpressed Gene Module Detection Algorithm Using Spectral Clustering

Microarray-based gene coexpression analysis is widely used to investigate the regulation pattern of a group (or cluster) of genes in a specific phenotype condition. Recent approaches look for differential coexpression patterns, where there exists a significant change in coexpression pattern between two phenotype conditions. These changes happen due to the alternation in regulatory mechanism across different phenotype conditions. Here, we develop a novel algorithm DCoSpect to identify differentially coexpressed modules across two phenotype conditions. DCoSpect uses spectral clustering algorithm to cluster the differential coexpression network. The proposed method is assessed by comparing with state-of-the-art techniques. We show that DCoSpect outperforms the state of the art in terms of significance and interpretability of detected modules. The biological significance of the discovered modules is also investigated using GO and pathway enrichment analysis.