Research And Application Of Machine Learning In Identification Of Protein Phosphorylation Nitration And Sulfation Sites

Protein post-translational modification(PTM),also known as covalent modification,is an important mechanism for regulating protein function.It plays an important role in signaling pathways and biological processes,and reversibly determines the dynamics and plasticity of cells.However,with the high-throughput development of protein post-translational modification data,traditional experimental methods are often laborious,time-consuming and expensive,and quick and convenient prediction methods for calculating and identifying modification sites emerge at the right moment.In this dissertation,we established online prediction tools for tyrosine nitration,sulfation and phosphorylation,and the modification of phosphorylation based on serine,threonine and tyrosine sites in seven fungi,and carried out proteomic analysis on them.Details are as follows:1.Computational prediction and analysis for tyrosine PTMs via elastic net.The tyrosine residue has been identified as suffering three major PTMs including nitration,sulfation,and phosphorylation,which could be involved in different physiological and pathological processes.Therefore,it is of great significance and beneficial to predict nitration,sulfation and phosphorylation of tyrosine residues in the whole protein sequence.Here,we first make use of sequence profiles,physicochemical properties and evolutionary information to encode features.Sequentially,we introduce elastic net to perform feature selection and develop a predictor named TyrPred(http://computbiol.ncu.edu.cn/TyrPred)for predicting nitrotyrosine,sulfotyrosine,and kinase-specific tyrosine phosphorylation sites.Cross validation and independent test results show that the prediction performance can be improved significantly by using elastic network to extract important training features.We anticipate that TyrPred can provide useful complements to the existing approaches in this field.2.Computational prediction and analysis of species-specific fungi phosphorylation via feature optimization strategy.Protein phosphorylation primarily occurs at serine,threonine and tyrosine residues and regulates a variety of biological processes.At present,the computational prediction of eukaryotic protein phosphorylation sites is mainly focused on animals and plants,especially on human,with a less extent on fungi.Therefore,more attention has been paid on the identification of fungi-specific phosphorylation.Based on the collected experimental fungi phosphorylation sites data,most of the sites are classified into different types to be encoded with various features and train via a two-step feature optimization method.A novel method for prediction of species-specific fungi phosphorylation-PreSSFP is proposed(http://computbiol.ncu.edu.cn/PreSSFP).The results of motif and feature analyses exhibited that there have some significant differences among seven species,which can provide a new lead for future computational analysis of fungi phosphorylation.