Poplar Protein Secondary Structure And Phosphorylation Sites Prediction

As one of the fast growing tree species,poplar has an important position in the forestry industry and the national economy,and it is also a model tree species of woody plant bioinformatics research.The structure and function of poplar protein is a hot research topic in the post genomic era.Taking secondary structure and phosphorylation modification sites of poplar as an important component of forest resources;using the BP artificial neural network technology to predict and research,a unified,simple and intuitive prediction model is constructed.The scope of forest resources management object is extended from macro to micro,which provides a new way for the research and development of forest bio informatics.This study presents protein physicochemical properties of amino acid type V descriptors encoding method which based on protein physical and chemical properties;performance of the model which using V coding method have been improved.It’s a more appropriate protein coding method.The main contents and results of this study are as follows:(1)At present,the data of poplar protein structure is still mainly obtained by experimental method,which acquisition speed relatively slow.The research and development of protein structure is very large.Using ANN technology to predict the protein structure of poplar is conducive to speed up the speed of protein structure data acquisition,and promote the improvement of the level of information management of forest resources.Using MATLAB software to realize the change of text data to numerical model,model was stored by the aid of the super conventional simulation function,which is convenient for researchers to visit and invoke.(2)Protein sequences of poplar fragments is harvested the length of 21 amino acid residues by sliding window technique as input data and corresponding secondary structure as output data and using amino acid V type descriptor encoding method encoding,is constructed based on BP artificial neural network of poplar protein secondary structure prediction model.Best model structure is 21:55:7,on the overall fit of the model accuracy 84,66%,for the secondary structure of a single type of fitting accuracy can reach 90.43%;model prediction accuracy rate is 74.26%,for the prediction of the secondary structure of a single accurate rate maximum reaching 82.02%,compared with the previous studies fitting precision and prediction accuracy was improved;model has a strong predictive ability.(3)Taking the length of 21 amino acid residues of poplar protein phosphorylation modification sequence as input data,the central of the amino acid residues was phosphorylated modification or not as output data to construct the prediction model based on BP artificial neural network of poplar protein phosphorylation modification sites,the best model structure is 21 x 16:8:4 and model’s ACC,Sn and SP,MCC is 78%,89%,67%,0.57,in addition to the specificity index SP,other indicators is better than previous research,model has a strong predictive ability.(4)Phosphorylation of S and T residues occurred in the phosphorylation sites of protein phosphorylation,and the phosphorylation of Y residues was not found;this is a feature that is a difference between the phosphorylation of proteins and other organisms;There is a certain relationship between the secondary structure and phosphorylation modification of poplar protein,and the structure of H type is more likely to be modified by phosphorylation.B,G and S type is not easy to be modified by phosphorylation.To sum up the research content and results,in the post genomic era,taking biological information as a component of forest resources,managing forest resources by micro level,It is feasible to use ANN technology to study the biological information of poplar,which of great significance to accelerate the development of poplar bioinformatics and to improve the level of information management of forest resources.