Computational Identification And Functional Analysis Of Protein Lysine,Threonine Post-translational Modification

Protein post-translational modifications(PTM)indicate changes in the function and structure of proteome after biosynthesis.PTM proteins are closely linked to human diseases,including cancer,diabetes,metabolic syndrome and many neurological diseases.At present,the identification of PTM can be obtained through two methods,one is to use high-throughput experimental technology,and the other is to calculate the recognition method.The calculation method is based on the high-throughput experimental identification data.Then it can provide the experimental researchers with accurate prediction of protein PTM to make the experiment more targeted,thus shortening the experimental period and reducing the experimental cost.In this paper,a variety of PTM were predicted based on protein sequence,data optimization and multiple feature extraction methods.The main work is summarized as follows:1.Protein PTM is a very important cellular control mechanism that can change the physical and chemical properties of proteins.Protein PTM is divided into single type and multitype PTM.The multitype PTM refers to two or more modifications to an amino acid.This type of PTM can also be called co-modification.Because of the changing science and technology,researchers have found more and more protein PTM sites,and it will be important to use machine-learning methods to process these data.Not only that,but we also need to build a website platform that presents these data processing results in front of you.2.The online prediction tool GlutPred is developed.Protein glutarylation is a novel post-translational modification(PTM).Comparing with the labor-intensive and time-consuming experimental approaches,in silico prediction can provide an effective method to the identification of various types of PTM sites.However,to the best of our knowledge,there is still no online predictor for glutarylation.Moreover,as we know one amino acid can be encoded by more than one m RNA codon-triplet.So the mRNA sequence would contain more possibilities information for the residues around the central site than protein sequence.However,is this information related to the glutarylated PTM? Unfortunately,the relevant reports are not found by searching the literature.To build the online predictor and answer this question,the experimentally identified glutarylation sites from mouse tissue proteins are selected,and the NCBI is searched with the each protein name to manually get the corresponding mRNAs.A simple matrix which is suitable to these two types of sequence at the same time is performed and an algorithm is developed.A user-friendly web interface is now freely available at http://bioinfo.ncu.edu.cn/GlutPred.aspx.3.The mThrPred,an online prediction tool for multi-label threonine,is constructed by multi-feature support vector machine algorithm.Protein carbonylation,glycosylation and phosphorylation are the important protein post-translational modifications(PTM).This kind of multiplex threonine residues may have some special functions worthy that are noticed for both basic research and drug development.To our best knowledge,although there are many computational methods to identify various single-label PTM types at threonine,no online predictor has ever been developed for multi-label PTM types at threonine.We establish an online prediction tool,called mThrPred,which can predict carbonylation glycosylation and phosphorylation sites at threonine.The novel predictor is developed by integrating multifarious informative features.The martensite correlation coefficients of glycosylation,carbonylation and phosphorylation are 0.8060,0.9370 and 0.6876,respectively,through five-fold cross-validation.The characteristic analysis shows that our forecasting tools have the advantages of high accuracy and good robustness.A user-friendly web interface is now freely available at http://bioinfo.ncu.edu.cn/mThrPred.aspx.4.BASPred,an online prediction tool for acetylation,succinylation and their co-modification,is constructed.Protein acetylation and succinylation are the significant protein post-translational modifications(PTM)involved in the plants' regulation of development and growth.It has become a meaningful topic in bioinformatics to understand various biological processes and predict the post-translational modification(PTM)sites in proteins.To our best knowledge,although there are many computational methods to identify acetylation and succinylation sites on mammalian and microbial proteins,no online predictor has ever been developed for these modifications in plants.We establish an online prediction tool,called BASPred,which can predict acetylation and succinylation sites in Seedling Leaves of Brachypodium distachyon L..It represents the first plant's PTM predictor that can be used to predict these modifications.The novel predictor is developed via integrating multifarious informative features with the optimizing training dataset by the NSEA(Negative Set Enrichment Analysis)method.Meanwhile,the cross-talk modification between acetylation and succinylation is first considered alone.The martensite correlation coefficients of acetylation,succinylation and co-modification are 0.7142,0.6801 and 0.6788,respectively,through ten-fold cross-validation.The characteristic analysis shows that our forecasting tools have the advantages of high accuracy and good robustness.A user-friendly web interface is now freely available at http://bioinfo.ncu.edu.cn/BASPred.aspx.