Title: Automatic segmentation of Potyviridae family polyproteins

Authors: Jheyson Faride Vargas; Jairo Andrés Velasco; Gloria Inés Alvarez; Diego Luis Linares; Enrique Bravo

Addresses: Departamento de Electronica y Ciencias de la Computacion, Pontificia Universidad Javeriana, Cali, Calle 18, No. 118-250, Cali, Colombia ' Departamento de Electronica y Ciencias de la Computacion, Pontificia Universidad Javeriana, Cali, Calle 18, No. 118-250, Cali, Colombia ' Departamento de Electronica y Ciencias de la Computacion, Pontificia Universidad Javeriana, Cali, Calle 18, No. 118-250, Cali, Colombia ' Departamento de Electronica y Ciencias de la Computacion, Pontificia Universidad Javeriana, Cali, Calle 18, No. 118-250, Cali, Colombia ' Departamento de Biología, Universidad del Valle, Ciudad Universitaria Meléndez, Calle 13, No 100-00, Cali, Colombia

Abstract: We describe an automatic segmentation method for polyproteins of the viruses belonging to the Potyviridae family. It uses machine learning techniques in order to predict the cleavage site which define the segments in which said polyproteins are cut in their process of functional maturation. The segmentation application is publicly available for use on a website and it can be accessed through the web service interface too. The prediction models have an average sensitivity of 0.79 and a Matthews correlation coefficient average of 0.23. This method is capable of predicting correctly (coinciding with previously published segmentation) the segmentation of sequences which come from Potyvirus and Rymovirus, genera. However accurate prediction capabilities are affected when faced with either atypical sequences or viruses belonging to less common genera in the Potyviridae family. Future work will focus on establishing greater flexibility in this sense.

Keywords: automatic segmentation; Potyviridae family polyproteins; Potyvirus; hidden Markov models; HMMs; artificial neural networks; ANNs; bioinformatics; machine learning.

DOI: 10.1504/IJBRA.2015.073238

International Journal of Bioinformatics Research and Applications, 2015 Vol.11 No.6, pp.525 - 539

Received: 24 Jun 2014
Accepted: 02 Jun 2015
Published online: 29 Nov 2015 *

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