Identification Of Essential Proteins Based On Network Capital Assessment And Invalidating Protein Node

Essential proteins are the core material basis of living life activities, which are involved in almost all biological processes. They are important components in cells, transferring and performing physiological functions etc. In PPI networks, effectively identifying essential proteins contributes to discover activities mechanisms of macromolecules in cells and life activity patterns of cells. Moreover, it will be of great help to biotech pharmaceuticals and biological treatment. This paper proposes improved node importance evaluation criteria by analyzing the global and local characteristics of PPI network, and puts forward corresponding essential protein recognition algorithms. The main research work in this text includes:(1) The traditional elimination of protein node method has the problem of destroying network topology and easily leading to the non-connectivity of the network. In this paper, we define accessibility of protein nodes, and develop a new algorithm NCAIP which is based on the network capital assessment (NCA) and invalidating protein node. Under the promise of network topology integrity, algorithm NCAIP evaluates the importance of nodes and identifies essential proteins in Protein-Protein Interaction (PPI) network by analyzing the declining extent of network capital before and after invalidating protein node. The experimental results show that the invalidation of protein node method based on NCA identifies essential proteins faster than the elimination of protein node method based on NCA. Furthermore, in the Top1%and Top5%data, NCAIP is better than four typical Centrality Measures (Degree Centrality, Closeness Centrality, Betweenness Centrality, Eigenvector Centrality) and Weighted Clustering Coefficient on the identification of essential proteins and has high accuracy on the identification of essential proteins.(2) In PPI network, dependent relationship exists between protein nodes, especially between adjacent nodes. The process of enforcing and delivering biological functions by protein nodes is simulated to describe the importance degree of protein nodes. Meanwhile, we utilize the invalidation of protein node method in order to ensure the completeness of PPI network. Comprehensively, this paper develops a new algorithm NCCIP which is based on the network node correlation (NCC) and invalidating protein node, which is taking into account both local and global characteristics of PPI network. The experimental results show that NCCIP has a strong ability to identify essential proteins. Furthermore, in the Top10%, Top15%and Top20%data NCAIP outperforms four typical Measures (DC, CC, BC, EC and WCC) and the NCAIP algorithm.