Study On The Prediction Of Protein Eight-state Secondary Structure And Anticancer Peptide

The biological function of protein is largely determined by its spatial structure.So it is necessary to study the structure of the protein is the key steps of analyzing the function of the protein.In general,the prediction of protein secondary structure is an important step in the prediction of protein spatial structure.Protein secondary structure prediction is mainly protein 3-state secondary structure(alpha-helix,beta-strand,random coil)prediction,but compared to the protein 3-state secondary structure,protein 8-state secondary structure can provide more detailed structural information,and therefore more challenging,especially for proteins with less sequence homology.In this paper,we introduced a model for the prediction of protein eight-class secondary structure using quadratic discriminant algorithm(QDA)based on the feature combination.Firstly,we selected 200 proteins,in which the sequence identity is ranges from 25%to 30%.And we extracted the chemical shifts of six nuclei in 200 proteins as features.Secondly,we used these chemical shifts as features,and combined with hydrophilic-hydrophobic residues to predict the protein 8-state secondary structures.Finally,we achieved the overall prediction accuracy of 8-state secondary structures(Q8)80.7%in seven-fold cross-validation.In the same dataset,we compared with other prediction tools,such as C8-Scorpion online server,as well as support vector machine(SVM)and random forest(RF)algorithm.The results showed that our prediction model is superior to other prediction algorithms in terms of accuracy.The anticancer peptide is a kind of antimicrobial peptide which has obvious antitumor activity.The anticancer peptide can not only quickly and effectively eliminate pathogenic bacteria,but also can effectively act on human tumor cells.How to effectively dentify anticancer peptides is one of the hot issues in biomedical research in more than 10 years.Based on the published dataset,we add the three kinds of secondary structure information as a new feature in dataset.Then,we combined with 20 amino acids(20AAC)and 6 kinds of hydrophobic-hydrophilic amino acids(6HP)as features,and implemented the prediction using the quadratic discriminant method(QDA).Finally,we obtained the overall accuracy 86%in the 7-fold cross validation,when using three kinds of protein secondary structure compositions(3PSS)combined with six kinds of hydrophobic-hydrophilic amino acid compositions(6HP)as features.And the overall accuracy is 94%by using three kinds of protein secondary structure compositions(3PSS)combined with 20 amino acid compositions(20AAC)as features.The prediction results showed that the prediction accuracy is improved when adding the secondary structure information as feature.Moreover,we compared with other prediction work in the same dataset,showing that our model is superior to other algorithms.