Comprehensive Analysis Of Sequence Of Cell-penetrating Peptides Base On Geometric Methods

Cell penetrating peptides(CPPs)are a class of small molecule short peptides that can mediate endocytosis by receptor dependent or non-receptor dependent manner can carry different molecules across the cell membrane,which provide an effective carrier for drugs targeting.The traditional method of CPPs has large workload and high cost.Therefore,bioinformatics methods are used to develop predictive analysis algorithms for the activity of CPPs and CPPs,and provide theoretical guidance for the experimental study of CPPs.bioinformatics methods has become an important auxiliary means for the study of CPPs.However,because the sequence of CPPs is short,most of the traditional methods of sequence analysis are no longer applicable.Therefore,effective researches on the series of important issues such as sequence structure mining of CPPs,prediction of classification,and prediction and prediction of penetration efficiency have been conducted.Rarely,the transmembrane mechanism of CPPs is still controversial.In this paper,firstly,the secondary structure and sequence characteristics of different types of CPPs were analyzed in depth.Based on this,fusion amino acid classification and the physiochemical properties of amino acids,a new method of geometric analysis of protein sequences was proposed.The predictive algorithm for classification of CPPs and non-CPPs yielded better prediction results and provided new ideas for the prediction of CPPs in the future.The main work of the thesis is as follows:1.Comprehensive analysis of secondary structures of different types of CPPs.In this section,the data source of CPPs and their secondary structure information is the CPPsite2.0 database.At present,there are 355 CPPs with annotation of secondary structures in the latest version of the database,which 340 annotations have specific information of transport substance and 187 annotations have clear transmembrane approaches.In this paper,the secondary structures of different types of CPPs are comprehensively analyzed for the first time.The results show that the different types of cell penetrating peptides tend to form random structure,and demonstrate various properties at secondary structure level with different degrees,these properties can provide solid theoretical basis for later study of CPPs.2.CPPs prediction based on a novel spherical representati.In order to dig protein sequence information,we used the spherical coordinates and fuse amino acid classifications and physiochemical properties of amino acids to proposed a new method for protein sequence analysis.In combination with support vector machines,we developed a predictive classification method for CPPs.Its prediction efficiency in multiple datasets shows that this method can provide more protein sequence information and can provide an effective tool and a new idea for protein analysis in the future.