基于粗集的T细胞表位预测方法

T细胞表位预测技术对于减少实验合成重叠肽、研究病原体与机体作用的免疫机制以及深入理解T细胞介导的免疫特异性均有重要意义。为增强T细胞表位预测模型的可理解性,本文在通过肽的预处理构建出存储等长肽段的决策表之后,设计出了一种基于粗集的T细胞表位预测方法。该方法由基于信息熵的属性约简完备算法和基于锚点知识的属性值顺序约简改进算法共同组成。基于HLA-DR4（B10401）编码的MHCII类分子结合肽的实验数据表明,在预测精度与传统神经网络方法大致相当的基础上,本文方法可以提取出用于帮助专家理解MHC分子与抗原肽结合机理的产生式规则。

T Cell Epitope Prediction Approach Based on Rough Set Theory

Predicting which peptides can bind to a specific MHC molecule is indispensable to minimizing the number of peptides required to synthesize,to the research on the interaction mechanics between infector and organism,and especially to helping understand the specificity of T-cell mediated immunity. In order to enhance the understandability of existing T cell epitope prediction methods based on machine learning,we firstly construct a decision table comprising the nonamers by peptide preprocessing. And then we propose a T Cell epitope prediction approach based on rough set theory,which consists of the complete attribute reduction algorithm based on information entropy and the renovated version for orderly attribute value reduction algorithm combined with expert knowledge of binding motifs. Finally,with the help of the approach,a comprehensible rule set with strong generalization ability to predict the peptides that bind to HLA-DR4(B1*0401)is acquired.