Effective 3D protein structure prediction with local adjustment genetic-annealing algorithm.

The protein folding problem consists of predicting protein tertiary structure from a given amino acid sequence by minimizing the energy function. The protein folding structure prediction is computationally challenging and has been shown to be NP-hard problem when the 3D off-lattice AB model is employed. In this paper, the local adjustment genetic-annealing (LAGA) algorithm was used to search the ground state of 3D offlattice AB model for protein folding structure. The algorithm included an improved crossover strategy and an improved mutation strategy, where a local adjustment strategy was also used to enhance the searching ability. The experiments were carried out with the Fibonacci sequences. The experimental results demonstrate that the LAGA algorithm appears to have better performance and accuracy compared to the previous methods.