The Research Of Conformational B-cell Epitope Prediction Method Based On Information Fusion And Computational Intelligence

An antigen is a molecule capable of inducing an immune response,whichcauses an immune system to produce antibodies against it.In the specific reaction of antigen and antibody, the specific chemical group specific antigen is called epitope antigen.The epitope and the antibody cause immune response.A B-cell epitope is a group of residues which recognized by a particular B-cell receptor or a particular antibody molecule patch. A B-cell epitope plays an important role in the acquired immune antibody, which plays an instructional role for preparation of vaccine, prevention and treatment of disease.Therefore, in the recent years, the research on prediction of B-cell epitope has become a hot issue.From the mimotopes and antigen antibody binding specificity, the mimotopes and the 3D structure of the antigen, the attribute information of the antigen,this thesisuses information fusion and intelligent computing methods to study the B-cell epitope prediction algorithm.The main results are as follows:(1) This thesis proposes a new algorithm based on the prediction of conformationalB-cell epitope and the antibody antigen binding specificity.The algorithm use mimotope sequence information and the information of antibody to find B-cell epitope.The method not only considers antigen information, but also thinks of the specific antigen and antibody binding information.In the same test set, the performance of the new proposed method is greater than several only considering epitope information methods.(2) This thesis puts forward a new algorithm based on the mimotopes and the 3D structure informationof conformational B-cell epitopes. In the process of predication, first of all, according to the 3D structural information of antigens calculate antigen of surface amino acid residues information(These surface amino acid residues are more likely to become the epitope).Then calculate the mimotopes information.Through the combination of the two, the antigens on the surface of amino acid residues are no longer a simple superposition of amino acid residues. However, according to the mimotopes, the surface amino acid residues are divided into with statistical significance of amino acid residue region. Finally, in order to obtain more predictive epitope, these regions are combined.The reasonable combination of amino acid residues in the region are the prediction results of the methods.The sensitivity and accuracy of the algorithm have obvious improvement.(3) This thesiscomes up with a new algorithm based on the conformational epitopes of B-cell epitope attribute information fusion.In the process of predication, with the reference of the relevant attribute information to reflect the strength of the epitope, using the correlation algorithm of computational intelligence on the predicted amino acid residues are classified to predict epitope information and non-epitope information.The result shows that the prediction performance of the method is relatively stable.At the same time,it indicates that the attribute information of the epitopes prediction algorithm has certain contribution.However, the problem is how to explore the effective conformation of B-cell epitope attribute informationto improve the accuracy of prediction.(4) Based on the above algorithm,this thesis realizes the online prediction platform and provides online forecasting service.According to the different input conditions, it uses different forecasting methods in order to achieve the prediction of conformational B-cell epitopes.The users can use prediction platform to gain more accurate resultsaccording to different situations.The online prediction platformachieves the improvementsfrom conformational B-cell epitope theory research to practical applications.