Integrated network analysis for studying human lung squamous cell carcinoma

Background Lung carcinoma is one of the most frequently diagnosed cancers as well as the most common cause of cancer death worldwide. Thus, research groups of all over the world aim at improve the prevention, diagnosis, and treatment of this type of cancer. The study of oncogenic properties of tumor cells requires the combination of experimental data and computational approaches supporting better disease knowledge as well as diagnosis and prognosis. Among the computational methods, gene network analysis, taking into account relationships among genes, offers a powerful tool to identify key players that mark the shift from normal to cancer state. Methods We propose an integrated network-based approach to analyse the correlation network arising from large-scale gene expression data. Our approach represents a novel perspective to the problem of node classification by assigning roles to nodes on the basis of network topology together with the gene expression data. Our rationale is that the biological roles of nodes should be identifiable looking at local interactions of each node with respect to the global connectivity of the network. Results We applied our methodology to the RNA and microRNA-sequencing data of human lung squamous cell carcinoma obtained from The Cancer Genome Atlas (TCGA) repository for patients with cancer and matched-normal tissues for both RNA and microRNA data. Our integrated network analysis allowed to identify a small pool of genes (about hundreds) called "switch genes" that are no local hubs and mainly interact outside their community, and may represent putative key gene regulators of human lung squamous carcinoma. Specifically, we unveiled 274 switch genes (out of ~1700 differentially expressed genes) that are significantly enriched in cell cycle annotation, supporting their crucial role in normal to cancer state transition. In the near future, we intend to extend our approach to the analysis of all tumours provided by TCGA.