论文部分内容阅读
抗体分子对蛋白抗原的结合主要通过表位区域进行,而表位区域的氨基酸残基通常形成不连续的、构象的或者空间的表位区域,而不是抗原表面上的线性连续片段.已有许多算法可以用来预测构象表位,并且基于各自的测试集,各种工具对空间表位的预测都声称取得了杰出的效果.本文收集了由实验方法确定的空间表位数据并建立了一套独立的测试集.基于这套测试集,采用灵敏度、真阳性预测率、成功挑选率和接受者操作特性曲线下面积(AUC)等参数对常用蛋白抗原空间表位预测工具进行了评估,工具包括SEPPA,CEP,DiscoTope,ElliPro,PEPOP和BEpro等.测试集评估结果表明,6种蛋白抗原空间表位预测工具预测性能仍有待提高.其中,SEPPA预测性能相对较好,然而计算得到的灵敏度、真阳性预测率、成功挑选率和平均AUC值也并不理想.评估结果还表明,预测工具采用的空间表位训练和测试数据集以及预测算法对预测结果的准确性有较大影响.以上分析结果为改进B细胞蛋白抗原空间表位预测方法和为免疫原性多肽和新型疫苗分子的设计提供新的启示.
Binding of antibody molecules to protein antigens occurs primarily through epitope regions, whereas amino acid residues in epitope regions typically form discrete, conformational or spatial epitope regions rather than linear contiguous fragments on the surface of the antigen. Many Algorithms can be used to predict conformational epitopes, and based on their respective test sets, various tools are claimed to have achieved outstanding results in predicting spatial epitopes.This paper collects the spatial epitope data determined by experimental methods and establishes a set of Independent Test Set Based on this set of tests, commonly used protein epitope prediction tools were evaluated using sensitivity, true positive predictive power, success selection rate, and area under the receiver operating characteristic curve (AUC) SEPPA, CEP, DiscoTope, ElliPro, PEPOP and BEpro, etc. The evaluation results of the test set show that the predictive performance of the six protein epitope prediction tools still needs to be improved, of which SEPPA has relatively good predictive performance. However, the calculated sensitivity The positive predictive value, the successful selection rate and the average AUC value are also not ideal.The evaluation results also show that the prediction tools used in spatial epitope training and Test data set and the prediction accuracy of the prediction algorithm greatly affect the results. In order to improve the results of the above analysis of B cell epitopes of protein antigens spatial prediction method and Implications for the design of new and novel vaccines immunogenic polypeptide molecule.