A novel and efficient method for differential pathway identification

Identifying differential pathways plays important roles in deciphering phenotype-specific cellular mechanisms. In this paper, we propose to decompose pathways into gene chains for detecting pathway expression changes. The decomposition into gene chains allows capturing local information on differential expression of pathways. To avoid biases of single gene set test methods, we collectively used multiple gene set tests to test whether a gene chain is significantly differentially expressed between two phenotypes. Focusing on p53 signal pathway that is well-recognized to be cancer-related, we evaluated the proposed method on two public available gene expression data sets, Liver cancer and Leukemia. Experimental results show the effectiveness and efficiency of the proposed method in identifying differentially expressed pathways.