Analyzing updates in Amino Acid Composition and Translation Algorithm towards Predicting Membrane Proteins using Machine Learning Approaches

Abstract

Membrane proteins are of different types that take on different functions. Classification of protein sequences in a data set is very important for understanding cell functions, disease prevention, and drug discovery. Initially, traditional methods were used for transmembrane protein classification. However, due to advanced technology and new research, it increases the transmembrane protein datasets by thousands which are almost impossible to obtain accurate results based on traditional methods. Computational methods are very useful for membrane protein classification. Several methods such as Pseudo Amino Acid Composition (PseAAC) can extract many silent features of a protein sequence. In this work, we intended to modify an existing algorithm of amino acid composition and translation to extract membrane protein features with better accuracy. To validate our algorithm, we will use the Support Vector Machine SVM and KNN.