Classification of Protein 3D Structures Using Artificial Neural Network

Abstract

Three dimensional (3D) protein structures determine the function of a protein within a cell. Classification of 3D structure of proteins is therefore crucial to inferring protein functional information as well as the evolution of interactions between proteins. In this paper, we utilized the artificial neural network (ANN) paradigm to classify the protein structures. The approach equally divides a 3D protein structure into several parts and then extracts statistical features from each part. The different parts in spatial domain are thus viewed as the feature vectors which are input into the neural network. The computational experiments achieve better results in most cases than other currently existing approaches. Index Terms—Protein structure; classification; ANN.