基于粗集的T细胞表位预测方法  

T Cell Epitope Prediction Approach Based on Rough Set Theory

作　　者：曾安[1] 潘丹[2] 郑启伦[3] 彭宏[3] 

机构地区：[1]广东工业大学计算机学院,广州510006 [2]中国移动通信集团广东有限公司,广州510100 [3]华南理工大学计算机科学与工程学院,广州510640

出　　处：《计算机科学》2007年第6期226-230,共5页Computer Science

基　　金：国家自然科学基金重点项目(No.30230350);广东工业大学博士启动基金项目(No.063001)。

摘　　要：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。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 theo- ry, 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.

关 键 词：T细胞表位预测 粗集 规则获取 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]