Protein Structure And Function Of Research Methods

The prediction of protein structures and functions is an important task in the ear of post-genome. Concerning the problem, some researches are made in this dissertation. The main work of this paper can be summarized as follows:1. The objective function in the model of prediction of protein structure is a potential energy function, either physics-based potentials or statistic-based potentials. The united residue force field, which is brought forward by Baker Laboratory of Chemistry, Cornell University, is analyzed. Based on this research, nine optimization models are constructed, and these models consist of different energy terms; therefore, they can be used to evaluate the impacts of the terms in the prediction of protein structure.2. Considering that the objective function of the prediction models is of multiple variables and multiple minima, simulated annealing algorithm is used to solve the problem. An improved simulated annealing algorithm for continuous optimization problem is improved. The plentiful numerical experiments demonstrate the convergence and it's better for solving large-scale continuous global optimization problems. The nine models of Met-enkephalin are computed, and the impacts of the different terms in the prediction of protein structure are narrated.3. Lennard-Jones clusters is a simple yet reasonably accurate mathematical model of a real physical system. The global minima of this models can help us understand the complicated molecule conformation problem, and it's an important part in models of protein folding, then the solving of L-J problem is helpful for the prediction of protein structure. Considering the characteristic of L-J problem, we add the local minimum to the improved simulated annealing algorithm. The new algorithm reduces the computing time and enhances the rate of success. In addition, its convergence properties is proved, and we get the conclusion that when t - 0, the improved simulated annealing algorithm asymptotically converges to the global optimal solution in probability one.4. The prediction of protein function is also the important problem of ear of post-genome. Though the research for protein function started late, some progenies have been acquired. This paper gives a brief review on the methods of protein function studying.