Protein Secondary Structure Prediction Modeling Research Based On Neural Network

Artificial neural networks (ANNs) has so many advantages such as parallel-operation, distributed-memory, self-adaptability, error-permissibility and so on. So it is suitable for solving those hydrology problems that are hard to build model accurately and have high non-linear and many uncertain puzzles. Now it is quite adept at forecasting & prediction, pattern recognition and control application. The accomplishment of the Human Genome Project(HGP) ,and the completion of more other genomes, ANN will play bigger roles in computational biology and bioinformatics. ANN Model has been used for the structure prediction and analyses of nucleotide sequence and protein sequence .for example, it was used to predict the transcribe sites of prokaryotic and the recognition of the ribosome binding site of bacterium coli and so on.During of the life action, different proteins implement different biological function. The proteins function is decided not only by its primary structure, but also by its three-dimensional structure. There are closely re(?)ationship between protein's structure and its biological function. Inorder to elucidate the essential of vital phenomena and molecular pathogenesis, it is veryimportant to research their relationship. While the research of protein secondary structure is vitalin ascertaining its tertiary structure and researching its biological function. So the forecasting of protein secondary structure question is one of the important issue in bioinformatics. Nowadays there are series of research results about protein secondary structure, but it is not enough to obtain well prediction accuracy, and there is on answer about the theoretical calculation of the tertiary structure.Used the ANN's advantages of parallel-operation and self-adaptability, the paper carried through the forecasting of protein secondary structure. The major research work is outlined as follows.(1) At first, the paper distinguished particularly the different model methods about ANN, analyzed its mechanism and ascertained the proper neural network model—CPN model which was be used to predict the protein secondary structure. Then the paper built up the CPN structure frame, designed its study algorithm and its training algorithm.(2) Firstly, the paper analyzed the different factors of 20 amino acid which influence the protein secondary structure, made use of these factors coded every amino acid. Then author searched the several decade proteins from the PDB protein structure bank, coded these proteins based on the amino acid code and sequence order, and made up of the input variable and the aim variable of CPN.(3) Author programmed the study program and training program of the CPN network, made the training experimentation and emulational experimentation on the MATLAB interface, then...