MHC多肽结合位点鉴定及超类型识别

项目名称： MHC多肽结合位点鉴定及超类型识别

项目编号： No.61502292

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 计算机科学学科

项目作者： 沈文君

作者单位： 汕头大学

项目金额： 20万元

中文摘要： 人类白细胞抗原（HLA）是人类的MHC。主要功能是将多肽递呈到细胞表面以供T细胞识别。识别抗原肽对开发表位疫苗起到重要的作用，很多研究提出了基于生物和计算机手段识别抗原肽的方法，然而HLA基因的高度多态性限制了高人口覆盖率表位疫苗的设计。虽然每个HLA分子只能与特定的多肽集群结合，但观察发现不同HLA可以结合的多肽集群是重合的，因此MHC超类型提出的目的在于将MHC分子进行归类，从而应对繁复的HLA的多态性。鉴于此，在本项目中我们将致力于提出高性能的算法用于鉴定MHC多肽结合位点和识别MHC超类型。拟提出的新方法将建立在对MHC多肽结合性质、计量多维尺度分析及Lasso特征提取的基础上。我们将尝试把计量多维尺度分析应用于构建MHC抗原肽结合图谱，同时结合Lasso提取出抗原肽MHC结合的关键性位点。本项目的预期成果将在理论和应用两方面促进MHC抗原肽结合位点鉴定和MHC超类型识别的发展。

中文关键词： MHC多肽结合位点鉴定；MHC超类型识别

英文摘要： Human Leukocyte Antigen (HLA) is the human MHC. The main function of MHC molecules is to display antigenic peptides to the surface of cell where they will be recognized by T cells. Identification of antigenic peptides plays a vital role in developing epitope vaccines, thus a number of experimental and computational approaches have been developed for their identification. The HLA is highly polymorphic which limits the design of high population coverage epitope vaccines.Each HLA molecule only binds to specific peptides, but it has been observed that different HLA molecules have overlapping peptide binding sets and there have been several attempts to group them into supertypes accordingly.In view of this, we will lay our focus on developing novel high performance methods for MHC significant residues identification and supertype classification. Specifically, the newly proposed methods will be based on the analysis of MHC-peptide binding specificity and the theory of metric MDS analysis and Lasso feature selection. In this project, we will attempt to apply metric MDS appropriately in the construction of MHC-peptide binding map and then integrate it with Lasso algorithm for MHC significant residues identification. The anticipated output of this project will benefit MHC significant residues identification and supertype classification both in theory and in application.

英文关键词： MHC Significant Residues Identification;MHC Supertype Classification