Application of neural networks for protein sequence classification

Abstract

Protein sequence classijkation is modelled as a binaty classijkation problem where an unlabeled protein sequence is checked to see if it belongs to a known set of protein superfamilies or not. In this paper we used multilayer perceptrons with supervised learning algorithm to learn the binary class$cation. The training data consists of two sets -a positive set belonging to an identifed set of protein super,family and a negative set comprising sequences from other superfamilies. When applying neural networks the first problem to be addressed is feature extraction. In this paper we used the new feature extraction techniques proposed by Wang et al. [4]. Simulations reveal that the neural network is able to classifi with good precision for Myosin and Photochrome superfamilies in the data set that we have chosen as positive . Also the results for Globin superfamily are good, thus validating the methodology of feature extraction and the application of neural networks for protein sequence classijkation as suggested by Wang et al. But, for Actin and Ribonuclease superfamilies the network showed poor performance. One possible reason for this may be that the choice of sequences in the negative data set is not optimal. We conclude from this work that the classification performance depends upon a proper selection of sequences for positive and negative data sets.

Computational Technique for an Efficient Classification of Protein Sequences With Distance‐Based Sequence Encoding Algorithm

“…In this work, the MLP utilizes a feed‐forward neural network method to map the input data to a set of suitable outputs. The MLP neural network consists of three layers, input, hidden, and output layer (Sharma et al, ; Weinert and Lopes, ). One hidden layer is enough to deal with this kind of protein sequence classification problems.…”

“…In the previous work, the encoding methods such as n‐gram, amino acid compositions, motifs, physicochemical properties, position specific scoring matrix (PSSM), and so on were employed, but these methods fail to achieve the rich statistical characteristics and sequence order information (Wang et al ; Anastasiadis, Magoulas, and Liu ; Sharma et al ; Zhao et al ; Bandyopadhyay ; Blekas, Fotiadis, and Likas ; Zainuddin and Kumar ; Datta et al ; Hong et al ; Mansoori et al ; Saidi et al ; Vipsita et al ; Li et al ; Gupta, Niyogi, and Misra ; Srinivasan et al ; Liu et al ).…”