Annealing genetic algorithm for protein folding simulations in the 3D HP model

The protein folding problem, i.e., the prediction of the tertiary structures of protein molecules from their amino acid sequences is one of the most important problems in computational biology. This problem has been widely studied under the HP model in which each amino acid is classified, based on its hydrophobicity, as a hydrophobic (H) residue or a hydrophilic (or polar, P) one. The protein folding problem in the HP model is in fact to find conformations with the lowest energies for some benchmark amino acid sequences. A genetic algorithm (GA) is used to find the lowest energy conformation in this paper. Each time for a newly produced offspring individual, which is originated from the selection, crossover and mutation operator of two parent individuals, we adopt the new acceptance criteria based on the annealing strategy to let it pass into the next generation, and propose a so-called annealing genetic algorithm (aGA) to predict efficiently the protein folding conformations in the three-dimensional (3D) HP model. Eleven benchmarks are tested to verify the effectiveness of the proposed approach and the computational results show that aGA explores the conformation surfaces more efficiently than other methods, and finds new lower energies in several cases, which means that aGA is an efficient tool for the protein folding simulations.