Network-based survival analysis methods for pathway detection in cancer

We compare three penalized Cox regression methods for high-dimensional survival data in order to identify the pathways involved into cancer occurrence and pro- gression. We analyze each method with three gene expression datasets including breast, lung and ovarian cancer. More precisely, we focus on cancer survival prediction and on top signature genes. The goal of this study is to gain a deeper insight of the benefits and drawbacks of the regression techniques in order to find the pathways involved in a specific type of cancer and identify cancer biomarkers useful for prognosis, diagnosis and treatment.