Construction and analysis of protein-protein interaction networks based on proteomics data of prostate cancer

Currently, using human prostate cancer (PCa) tissue samples to conduct proteomics research has generated a large amount of data; however, only a very small amount has been thoroughly investigated. In this study, we manually carried out the mining of the full text of proteomics literature that involved comparisons between PCa and normal or benign tissue and identified 41 differentially expressed proteins verified or reported more than 2 times from different research studies. We regarded these proteins as seed proteins to construct a protein-protein interaction (PPI) network. The extended network included one giant network, which consisted of 1,264 nodes connected via 1,744 edges, and 3 small separate components. The backbone network was then constructed, which was derived from key nodes and the subnetwork consisting of the shortest path between seed proteins. Topological analyses of these networks were conducted to identify proteins essential for the genesis of PCa. Solute carrier family 2 (facilitated glucose transporter), member 4 (SLC2A4) had the highest closeness centrality located in the center of each network, and the highest betweenness centrality and largest degree in the backbone network. Tubulin, beta 2C (TUBB2C) had the largest degree in the giant network and subnetwork. In addition, using module analysis of the whole PPI network, we obtained a densely connected region. Functional annotation indicated that the Ras protein signal transduction biological process, mitogen-activated protein kinase (MAPK), neurotrophin and the gonadotropin-releasing hormone (GnRH) signaling pathway may play an important role in the genesis and development of PCa. Further investigation of the SLC2A4, TUBB2C proteins, and these biological processes and pathways may therefore provide a potential target for the diagnosis and treatment of PCa.