An evolutionary multi-agent algorithm to explore the high degree of selectivity in three-dimensional protein structures

Correa, Leonardo de Lima; Inostroza-Ponta, Mario; Dorn, Marcio; IEEE

Abstract

Tertiary protein structure prediction in silico is one of the most challenging problems in Structural Bioinformatics. The challenge arises due to the combinatorial explosion of plausible shapes, where a long amino acid chain ends up in one out of a vast number of three-dimensional conformations. The rules that govern the biological process are partially known, which difficult the development of robust prediction methods. Many computational methods and strategies were proposed over the last decades. Nevertheless, the problem remains open. The agent-based paradigm has been shown a useful technique for the applications that have repetitive and time-consuming activities, knowledge share and management, such as the integration of different knowledge sources and modeling of complex biological systems. In this paper, we propose a first principle method with database information for the 3-D protein structure prediction problem. We do so by designing a multi-agent approach that uses concepts of evolutionary algorithms to speed up the search phase by improving local candidate solutions from the protein conformational space. To validate our method, we tested our computational strategy on a test bed of eight protein sequences. Predicted structures were analyzed regarding root-mean-square deviation, global distance total score test and secondary structure arrangement. The obtained results were topologically compatible with their correspondent experimental structures, thus corroborating the effectiveness of our proposed method. As observed, the evolutionary multi-agent approach achieved good results in terms of the evaluted measures and was able to efficiently search the roughness of protein energy landscape.

Más información

Título según WOS: An evolutionary multi-agent algorithm to explore the high degree of selectivity in three-dimensional protein structures
Título de la Revista: 2017 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC)
Editorial: IEEE
Fecha de publicación: 2017
Página de inicio: 1111
Página final: 1118
Notas: ISI