| In natural protein, the final 3D space structure is decided by amino acid——primary structure exclusively, and protein 3D space structure largely determines its biological functions. Researchers also find that a lot of diseases of human are connected with protein mutation, so the acquaintance with protein space structure is especially important.Presently, there are two methods to predict protein structure which are experiment measurement and theory prediction. Experiment measurement method not only consumes time, but also is constrained by experiment conditions, and with the in-depth study, experiment measurement can't meet the study needs increasingly. With the development of biology technology and computer technology, theory prediction is gradually developed the ideal choice of protein structure prediction.More and more study results indicate that protein structure prediction is NP problem which can't be solved efficiently by a general method. In the event that we don't know general regulations of natural protein folding, to solve the problem many scholars present simplified models which become a tool to research basic properties of protein folding.This paper introduces a new method to the treatment of new key border values on the basis of the traditional simulated annealing algorithm, and applies to 2D AB off-lattice model. Firstly, we take test with four Fibonacci sequences for structure prediction and get better results which indicate that this algorithm is feasible and effective for protein structure prediction. Secondly, we apply this method to two authentic protein sequences for structure prediction simulation. The results indicate that the minimum values of energy we get are better than the results traditional simulated annealing algorithm gets. In addition, this paper also achieved another simulated annealing algorithm which improved by Zhang Hongjuan, applies it to protein structure prediction. The results show that the algorithm of this paper is better. |
PDF Full Text Request