Identifying essential proteins using modified-monkey algorithm (MMA). (October 2020)

• Record Type: Journal Article
• Authors: Payra, Anjan Kumar; Ghosh, Anupam
• Is Part Of: Computational biology and chemistry. Volume 88(2020)
• Publication Date: 2020-10
• Subjects: Classifier -- Hybrid weighted approach -- Complex density score -- Centrality; Chemistry -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; Biochemistry -- Data processing; Biology -- Data processing; Molecular biology -- Data processing; Periodicals; Electronic journals; 542.85
• DOI: 10.1016/j.compbiolchem.2020.107324
• Languages: English
• ISSNs: 1476-9271

Abstract

Graphical abstract: Highlights: We propose MMA that is a combined approach of climb, watch-jump and somersault sub processes to predict essential proteins of a PPIN. ECV and ECC approaches are predicted for directed-weighted and undirected network respectively in climb sub process. CDS is introduced to obtain highly connected and reliable proteins. Select centralities using classifiers in watch-jump sub process. NC, DC etc. are used to predict mostly connected set of proteins locally and EC predicts other essential proteins throughout the network. 3-sigma threshold based essential protein seeds are selected out of given PPIN throughout MMA. Unbiased centralities are nominated depending on classifier based benchmark approach in watch-jump sub-process. The proposed methodology has achieved higher success rates in comparison to the existing methods. Abstract: Proteins are the most essential macromolecules needed for the normal flow of life. Essential proteins play a key role to control other proteins in an interaction network for the growth and understanding of the molecular mechanism of cellular life. Though there are many computational algorithms for essential drug discovery depending on nature of essential proteins, but still lots of improvements and optimizations are required. In this work, modified-Monkey algorithm (MMA) is proposed for the identification of essential proteins in protein-protein interaction network (PPIN). Nature of a monkey can be distinctly described in three processes like climb, watch-jump, and somersault in different problem spaces. These processes of monkey traversal are plotted in PPIN with objective to find out essential proteins. Performance of MMA is assessed with other existing essential protein prediction methodologies, including Eigenvector Centrality (EC), Betweenness Centrality (BC), Network Centrality (NC) and others also. The proposed methodology has achieved higher success rates in comparison to the existing state-of-art model.