A New Method for Identification of Essential Proteins by Information Entropy of Protein Complex and Subcellular Localization

摘要

Essential proteins are critical components of living organisms. The identification of essential proteins from protein-protein interaction (PPI) networks is beneficial for the understanding of biology mechanism. This work presents a novel information entropy of protein complex and subcellular localization based method (IECS) for essential protein identification from PPI networks. First, extract the sample by stratified sampling to calculate the information gain of the protein complex and subcellular localization. Information gain can effectively determine the importance of biological characteristics. Then calculate the biological attribute score based on the information entropy of protein complex and subcellular localization. Finally combined with the network characteristics of the node. The proposed IECS method is implemented on two Saccharomyces cerevisiac datasets (DIP and Krogan), and the experimental results show that IECS overmatches most of the traditional methods for identifying essential proteins.