Abstract

BEDOYA, Oscar F.. Remote-3DD: a new remote homology detection method that uses physicochemical properties. Ing. compet. [online]. 2017, vol.19, n.2, pp.25-35. ISSN 0123-3033.  https://doi.org/10.25100/iyc.v19i2.5289.

In this paper, we present a new method for remote homology detection called remote-3DD. The proposed method combines predicted contact maps and distributions of the interaction matrices. Predicted contact maps approximate the 3D shape of a protein based on its primary structure. On the other hand, an interaction matrix allows representing a protein by using the physicochemical properties of its amino acids. The remote-3DD method is proposed as a strategy to improve the accuracy of the remote-C3D method, which uses contact maps alone. In this paper, we hypothesize that we can improve the accuracy of the remote-C3D method by including physicochemical properties. The results show that the accuracy of the remote-3DD method is higher than the accuracy of the composition-based methods and in some cases comparable with the accuracy of the profile-based methods. In addition, the results also show that the remote-3DD method achieves higher accuracy values than the remote-C3D method when considering the same number of models and size of submatrices.

Keywords : Bioinformatics; classifiers; physicochemical properties; remote homologs; SCOP dataset..

        · abstract in Spanish     · text in English     · English ( pdf )