A Novel Conformational B-cell Epitope Prediction Method Based On Patch And Amino Acid Pairs

The body will produce humoral immune responses when it is subjected to the stimulation from external antigen, a part of protein antigen recognized by a particular BCR is defined to a B-cell epitope. Locating B-cell epitopes is important for the purpose of effective vaccine design. Predicting epitopes by experimental methods is expensive in terms of time, cost and effort; so computational methods with its low cost and high speed are employed to predict B-cell epitopes widely. In recent years, epitope prediction based on mimotope has become a focus as a promising method. Some novel software and web-based servers have been proposed and have succeeded in given test cases.In this paper, we proposed a new epitope prediction method based on amino acid pairs and patches. Firstly,this algorithm divided the mimotope which got through screening the phage peptide library into amino acid pairs. Secondly,the amino acids on the surface of antigen were divided into overlapping patches, then the amino acids in each patch divided into pairs though a parameter also. Finally,this algorithm match the two kinds of amino acid pairs and got the optimal patch,the amino acids in it would be as the candidate epitope.In this work,we selected 39 data as the test data set from the benchmark data set Benchmark 2.0. On the test data set, the proposed method gave out the mean sensitivity of 0.53, the specificity of 0.77, the ACC of 0.75 and the F-measure of 0.45 on 39 test cases. Through comparing with mimotope-based methods, patch-based methods and two other methods, the sensitivity of our new method got a significantly improvement. The results demonstrated the proposed method is successful in the patch and AAPs analysis and competent for the conformational B-cell epitope prediction.