High-throughput identification of antibody features for sequence-based epitope prediction

点击翻译按钮获取中文摘要

PROJECT SUMMARY Human antibody repertoire is highly diverse due to VDJ recombination and somatic hypermutation. VDJ recombination is a somatic recombination process that assembles the variable region of an antibody from a diverse set of gene segments, known as variable (V), diversity (D), and joining (J) genes. During the course of an immune response, antibodies will increase affinity to their antigens through somatic hypermutation. The huge diversity of antibodies enables human immune system to confer protection against various pathogens by recognizing a wide range of antigens and epitopes. Detailed molecular characterization of antibody-antigen interaction is crucial to vaccine and therapeutic development, as well as the fundamental understanding of the human immune system. The binding specificity and epitope of an antibody are determined by its structure, which in turn is determined by its amino acid sequence. As a result, information on the binding specificity and epitope of an antibody are encoded in its amino acid sequence. However, accurately predicting the epitopes of antibodies from their sequences is an extremely difficult task because our understanding of antibody sequence-function relationship is far from comprehensive. This proposal aims to develop a library-to-library screening approach to characterize antibody-antigen interaction in a high-throughput manner, with a focus on influenza A hemagglutinin (HA) as a proof-of-concept. Specifically, we will determine the HA-binding specificity and conformational epitope of hundreds of thousands of antibodies in a single experiment. Subsequently, antibody sequence features that are associated with different epitopes on HA will be systematically identified. We further aim to use these antibody features to identify HA-binding antibodies from publicly available antibody repertoire sequencing datasets as well as predict their epitopes on HA. While this proposed project focuses on influenza HA, our approach can be easily extended to any antigen of interest. This proposal will open up the possibility for antibody sequence-based epitope prediction and provides new perspectives to the understanding of human antibody repertoire.