A genetic algorithm with backtracking for protein structure prediction

Abstract

In this paper, we propose a simple genetic algorithm for finding the optimal conformation of a protein using the three-dimensional square HP model. A backtracking procedure is used to resolve the positional collisions and illegal conformations that occur during the course of genetic search. Backtracking is shown to be a simple and efficient means of collision repair that requires little overhead. Empirical results show that a genetic algorithm using backtracking can obtain the lowest energy structure of an amino acid sequence in fewer energy evaluations than earlier approaches.

Enhanced Genetic Algorithm for Protein Structure Prediction based on the HP Model

“…Tables 3 and 4 show the results of our EGA in comparison with the previously published results of Johnson et al (2006) for proteins with lengths 27 and the results of Patton et al (1995) for 64 amino acids, respectively . Johnson et al (2006) in 9 out of 10 cases with a better value in the remaining case.…”

“…Tables 3 and 4 show the results of our EGA in comparison with the previously published results of Johnson et al (2006) for proteins with lengths 27 and the results of Patton et al (1995) for 64 amino acids, respectively . Johnson et al (2006) in 9 out of 10 cases with a better value in the remaining case. However, the numbers of evaluations of fitness values is significantly less in the afore-mentioned 9 cases; it is greater in the 10 th case where EGA clearly managed to search larger areas of the search space that enabled it to find a fitter protein structure.…”

“…In this section, we report the empirical results of the proposed GA and compare them to those of published techniques: one by Patton et al (1995), which proposed a standard genetic algorithm for this problem and reported better results than those achieved by Unger and Moult (1993b); the second is by Johnson et al (2006), which reports better results than those achieved by Patton et al (1995) for the smaller sequences. We use two sets of benchmark sequences employed first by Unger and Moult (1993b).…”

Algoritmos evolutivos e modelo HP para predição de estruturas de proteínas

“…O AE implementado foi comparado com outros trabalhos encontrados na literatura (Patton et al, 1995;Johnson and Katikireddy, 2006;Custódio, 2008). Primeiramente, validouse o modelo, comparando-o a um algoritmo genético (AG) padrão (Patton et al, 1995), utilizado como referência em diversos trabalhos.…”

“…Posteriormente, comparou-se a abordagem proposta com duas abordagens não tradicionais e mais recentes. A primeira (Johnson and Katikireddy, 2006) consiste em um AG híbrido e a segunda é uma hiper-heurística (Custódio, 2008). Resultados expostos neste artigo motivam trabalhos futuros combinando o AEMT proposto para PSP com aspectos de AG híbrido e de hiperheurísticas.…”