Article: Evaluating biological characteristics for protein function prediction using support vector machine Journal: International Journal of Bioinformatics Research and Applications (IJBRA) 2021 Vol.17 No.1 pp.1 - 24 Abstract: Predicting protein function is a latent problem and a challenge in the field of Bioinformatics. Over the years, several computational approaches have been proposed for this purpose. One of the approaches is based on characteristics, which makes use of biologic relevant information. The several contributions with this approach have considered a set of characteristics belonging to the four protein structures to predict the class and function of proteins. In this study, we evaluate several sets of characteristics that represent the four structural levels (primary, secondary, tertiary and quaternary), such as electrostatic potential, hydrophobicity, amino acids frequency, distance between α-carbons, and molecular weight. In order to evaluate the relevance of the characteristics, we employed the support vector machine (SVM) classifier, which usually presents satisfactory results in the process of biological data classification. The objective of this study is to contribute for the most appropriate choice of characteristics for the protein function prediction problem. Inderscience Publishers - linking academia, business and industry through research

Title: Evaluating biological characteristics for protein function prediction using support vector machine

Authors: Gabriela Teodoro De Oliveira Santos; Cristiane Neri Nobre; Luis Enrique Zárate

Addresses: Computer Science Department, Pontifical Catholic University of Minas Gerais, Belo Horizonte, Minas Gerais, 30535-901, Brazil ' Computer Science Department, Pontifical Catholic University of Minas Gerais, Belo Horizonte, Minas Gerais, 30535-901, Brazil ' Computer Science Department, Pontifical Catholic University of Minas Gerais, Belo Horizonte, Minas Gerais, 30535-901, Brazil

Abstract: Predicting protein function is a latent problem and a challenge in the field of Bioinformatics. Over the years, several computational approaches have been proposed for this purpose. One of the approaches is based on characteristics, which makes use of biologic relevant information. The several contributions with this approach have considered a set of characteristics belonging to the four protein structures to predict the class and function of proteins. In this study, we evaluate several sets of characteristics that represent the four structural levels (primary, secondary, tertiary and quaternary), such as electrostatic potential, hydrophobicity, amino acids frequency, distance between α-carbons, and molecular weight. In order to evaluate the relevance of the characteristics, we employed the support vector machine (SVM) classifier, which usually presents satisfactory results in the process of biological data classification. The objective of this study is to contribute for the most appropriate choice of characteristics for the protein function prediction problem.

Keywords: prediction of protein function; biological characteristics; SVM; support vector machine.

DOI: 10.1504/IJBRA.2021.113961

International Journal of Bioinformatics Research and Applications, 2021 Vol.17 No.1, pp.1 - 24

Received: 07 Jun 2017
Accepted: 01 May 2018
Published online: 06 Apr 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Evaluating biological characteristics for protein function prediction using support vector machine

Keep up-to-date