The Research On The Protein Structure Prediction Based On The Off-lattice Protein Model

The function of protein is directly given by its advanced structure as a result of the folding of its polypepsides and the folding information contains in its amino acid sequence or its primary structure.So, the structure prediction from the primary structure to advanced structure becomes an important task in the protein research field. Concerning the problem, some researches are made in this dissertation.The main work is summarized as follows:1. Based on the protein 2-dimension HP off-lattice model and the improved simulated annealing algorithm , this paper discusses the respective roles of short- and long interactions in protein folding .the configurations of 1ECD,2RNS,1PHT,1WBC are obtained, and with those structure informations in the PDB , the influence of long-range interactions is discussed. The results show that for the folding and stability of proteins, long-range interactions are the main determinants.2. Evolution strategies as global optimization search methods, are computing models simulating the evolutionary mechanisms of nature. An improved evolution strategies is applied to protein 3-dimension HPNX off-lattice model, and we successfully predict the structures of several protein sequences, such as 1RPB 1BPI and 1UBQ. These results show that HPNX off-lattice model is more accurate than HP off-lattice model in representing the folding trends of proteins and evolution strategies are feasible and effective to the prediction of protein structures.