基于结合矩阵算法构建HLA超型特异性CTL表位预测及设计工具的研究

基于T细胞表位的多肽疫苗在病毒感染、自身免疫性疾病以及肿瘤的治疗中具有广阔的应用前景，然而由于人白细胞抗原(HLA)的高度多态性及其在不同种族人群中的分布差异，如何提高基于表位的多肽疫苗在人群中的覆盖率是疫苗设计的关键科学问题之一。超型特异性CTL表位因其可以和同超型不同HLA分子之间发生交叉反应，是开发具有高人群覆盖率疫苗的优先选择对象。本研究将选择合适的矩阵算法定量描述表位各个位点残基对其与同超型HLA分子交叉结合能力的贡献，建立基于互联网的超型特异性CTL表位的预测及设计工具；根据同超型HLA分子多肽结合槽的相似性扩展此工具的应用范围，建立可实现没有或者只有很少已知表位的HLA分子限制性表位的pan特异性预测方法。与现有表位预测工具进行比较，对其性能作出准确定位；以HBV相关抗原为研究对象，结合免疫学试验分别验证此工具在预测和设计超型CTL表位以及其pan特异性预测工具的效率。

The T cell epitope-based vaccines have the great potential for the development of epitope-based vaccines against viral infections, autoimmune diseases and even certain tumors. Because of the high polymorphisms of Human leukocyte antigen (HLA) alleles and the diversity of distribution in various populations, How to increase the coverage of epitope-based vaccines in human population is one of the critical scientific issues. Supertype-specific CTL epitopes which can bind to multiple HLA molecules within the same supertype have great potential for the development of vaccines with broad and unbiased coverage of the human population. In this study, we will select the most appropriate matrix-based algorithm to describe the quantitative contributions of residues at each position of the peptide to its cross-binding ability to HLA molecules, and establish a web-based tool for prediction and design of the supertype-specific CTL epitopes. Furthermore, we will generate the pan-specific method for the prediction of peptides bound to HLA molecules with only a small number of experimentally obtained data according the similarity of HLA-peptide binding grooves, which will expand the application of the tool for prediction and design of the supertype-specific CTL epitopes. The performance of this prediction tool will be compared with other online available tools,such as BIMAS, SYFEPITHI, SVMHC and netMHC; The efficiency of this web-based tool will be validated by identifying corresponding the epitopes of HBV related antigens using experimental methods.