G. Naveen Sundar and D. Narmadha* Pages 1 - 9 ( 9 )
Background: Essential proteins play a crucial role in most of the living organisms. The computer-based task of predicting essential proteins is important for target protein identification, disease treatment and suitable drug development.
Objective: Traditionally many experimental and centrality measures have been proposed by researchers to predict protein essentiality.
Methods: The prediction accuracy, sensitivity, specificity identified by the traditional methods is very low.
Result and Discussion: In this research work, a novel computational based approach such NC-KNN model has been proposed to identify the most essential proteins. The proposed work uses a combination of network topology measure and machine learning model to predict the essential proteins.
Conclusion: The proposed work shows a remarkable improvement than seven traditional centrality based measures such as DC, BC, CC, EC, NC, ECC and SC in terms of the metrics such as accuracy(A1), precision(P1), recall(R1), sensitivity(SE) and specificity(SP).
Graph connectivity, protein-protein Interaction (PPI), essential target protein, k-nearest neighbourhood.
Department of CSE, Karunya Institute of Technology and Sciences, Department of CSE, Karunya Institute of Technology and Sciences