Abstract:

To save the cost of experiment, improve the work efficiency, we predict the Bcell epitope by computing method, and then use the results to direct the experiment further. We extract 10 epitope relevant amino acids propensities, and then classify antigen surface residues by using vector machine algorithm to predict the candidate epitopes. According to fifteen test cases, we validate the effectiveness of the method.

Key words: epitope prediction, support vector machine(SVM), epitope features, classification