A Novel Method to Predict Essential Proteins Based on Diffusion Distance Networks

• Published in: IEEE access Vol. 8; pp. 29385 - 29394
• Main Authors: Zhao, Bihai, Han, Xiao, Liu, Xiner, Luo, Yingchun, Hu, Sai, Zhang, Zhihong, Wang, Lei
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: diffusion distance; Essential proteins; Gene expression; Graph theory; Mathematical model; Methods; Modularity; Network topology; Networks; Organisms; Prediction methods; protein-protein interaction; Proteins; Topology
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2020.2972922

Essential proteins are important for the survival and reproduction of organisms. Many computational methods have been proposed to identify essential proteins, due to the production of vast amounts of protein-protein interaction (PPI) data. It has been demonstrated that PPI networks have graph-theoretic characteristics as so-called small-world and scale-free. The traditional metrics cannot really reflect the relationship between proteins when identifying essential proteins from PPI networks. In this paper, we construct a diffusion distance network (DSN) by combining PPI topology characteristics with orthologous proteins and sub-cellular localization information of proteins. Taking the modularity feature of essential proteins into account, we proposed a new essential proteins prediction method based on DSN. We employed our DSN method and ten other state-of-the-art methods to predict essential proteins. The precision-recall curve, jackknife methodology and so on are used to test the performance of these methods. Experimental results show that our method outperform ten other competitive methods. The row data and the software are freely available at: https://github.com/husaiccsu/DSN.