Similarity measure for cell membrane fusion proteins identification

Abstract

This work proposes a similarity measure between secondary structures of proteins capable of fusing cell membranes and its implementation in a classification system. For the evaluation of the metric we used secondary structures estimated from amino acid sequences of Class I and Class II viral fusogens (VFs), as well as VFs precursor proteins. We evaluated three different classifiers based on k-Nearest Neighbors, Support Vector Machines and One-Class Support Vector Machines in different configurations. This is a first approach to the similarity measure with satisfactory results. It is possible that this method could allow the identification of unknown membrane fusion proteins in other biological models than the proposed in this work. Keywords: Cell Membrane Fusion, Viral Fusogen, Similarity Measure, Support Vector Machines, One-Class Support Vector Machines, k-Nearest Neighbors