A gene link-based method for identifying differential gene pathways

Pathway analysis plays an important role in exploring underlying connections between genomic data and complex diseases. In this paper, we propose a gene link-based method for identification of differentially expressed gene pathways. By viewing gene links in a pathway as a Markov chain, the proposed method first develops a gene link Markov chain model (MCM) and devises a Markov chain model-based classification rule to measure the biological importance of a gene link. Then, the expression difference of a pathway is estimated based on all the gene links in the pathway using the gene link MCM. The use of gene links, instead of individual genes, allows for exploring pathway topology that is crucial to pathway activity in cells. Results on two real-world gene expression data sets demonstrate that the effectiveness and efficiency of the proposed method in identifying differential gene pathways.