Protein Secondary Structure Prediction Based On Probability Graph Model HMM

Protein structure is closely related to protein function,and protein secondary structure is the basis of other high spatial structures formation,therefore,protein secondary structure prediction has become a hot topic in bioinformatics.Protein secondary structure prediction is to establish the model between amino acid sequences and the corresponding secondary structures based on the known protein secondary structure data,and through the model to predict the secondary structure of unknown amino acid sequences.Hidden Markov Model(HMM)is a probabilistic statistical model.Some scholars have applied it to the protein secondary structure prediction and have received some effects.In this paper,the 3-state HMM and the 7-state HMM are used to predict the protein secondary structure.The results show that the the latter HMM is better.With regard to the 7-state HMM,improve the method in structure state and parameter training.On one hand,Considering that the 7-state HMM does not take into account the state of the non-secondary structure,the state F is introduced.On the other hand,for the second underflow of the parameter revaluation process,we break the routine,without throughing some means to prevent its underflow,but to average the optimal revaluation parameters as the parameter of secondary structures prediction.Through these two improvements,the prediction accuracy is improved to a certain extent.The study shows that it is important to select the appropriate protein structure states and appropriate training set to improve the accuracy of the protein secondary structure prediction.