Computational methods for rapid structural modelling of antibody-antigen interactions to improve identification of antigen-specific antibodies from BCR-seq data
<p>Antibodies are immune proteins that are the basis of humoral immunity in jawed vertebrates, permitting highly-specific, mutable and lasting recognition of diverse foreign molecules. In their membrane-bound form, they are referred to as B cell receptors (BCRs): the collection of antibodies o...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2022
|
| _version_ | 1797112795099037696 |
|---|---|
| author | Richardson, E |
| author_facet | Richardson, E |
| author_sort | Richardson, E |
| collection | OXFORD |
| description | <p>Antibodies are immune proteins that are the basis of humoral immunity in jawed vertebrates, permitting highly-specific, mutable and lasting recognition of diverse foreign molecules. In their membrane-bound form, they are referred to as B cell receptors (BCRs): the collection of antibodies or BCRs in an individual is a record of their antigenic history, and tells us how their B cells have recognised and interacted with pathogens over their lifespan. Since the advent of next-generation sequencing and its application to these repertoires, we can sample and sequence the receptors of 103 to 106 B cells from a given individual. However, there remains an unmet need for mapping the mounting number of these BCR sequences to their putative antigen specificity. In this thesis, we aim to demonstrate how computational structural methods can be applied to BCR repertoires to improve our ability to identify antigen-specific antibodies.</p>
<p>In the first chapter, we describe a novel method for identifying antigenspecific antibodies from repertoire data using paratope prediction which we call "paratyping", and then use paratyping to discover novel Pertussis toxoidbinding antibodies from the BCR repertoires of transgenic mice.</p>
<p>Transgenic mice are a common source of therapeutic antibodies. To improve our understanding of the structural landscape of the BCR repertoires of these workhorses of antibody discovery, in the second chapter we apply structural annotation and modelling approaches to naive repertoires from humans, mice and transgenic mice. We show that the starting structural repertoires of transgenic mice are intermediate between humans and mice, despite being encoded by human genes, as a result of deficient junctional diversification in the primary repertoire of transgenic mice.</p>
<p>In the third chapter, we describe the creation of a novel sequence database for Ebolavirus-binding antibodies called EBOV-AbDab. Using this data alongside other techniques, we analysed the longitudinal BCR repertoires of 40 individuals vaccinated with Ad26.ZEBOV/MVA-BN-Filo. We used paratyping to predict the epitopes of the majority of the most expanded clonotypes after the booster vaccine, providing evidence of a highly convergent response to Ebolavirus vaccination that correlated with anti-Ebolavirus glycoprotein IgG titre.</p>
<p>Finally, in the fourth chapter we examine the growing body of nativelypaired data to try to understand VH:VL pairing preferences. We re-examine the VH:VL interface in the context of germline gene usage and subsequently analyse at scale the genetic pairing preference of antibodies in the largest paired sequencing repository, the Observed Antibody Space, in combination with 3x as much novel data produced by an alternative single-cell sequencing technology.</p>
<p>In the final chapter, I outline further work which can extend or complete the work presented in the previous chapters, as well as general outlooks for the field.</p> |
| first_indexed | 2024-03-07T08:04:51Z |
| format | Thesis |
| id | oxford-uuid:3da3cb7b-a42f-4223-b229-44ea2a50f273 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-04-09T03:54:30Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | oxford-uuid:3da3cb7b-a42f-4223-b229-44ea2a50f2732024-03-08T10:29:50ZComputational methods for rapid structural modelling of antibody-antigen interactions to improve identification of antigen-specific antibodies from BCR-seq dataThesishttp://purl.org/coar/resource_type/c_db06uuid:3da3cb7b-a42f-4223-b229-44ea2a50f273EnglishHyrax Deposit2022Richardson, E<p>Antibodies are immune proteins that are the basis of humoral immunity in jawed vertebrates, permitting highly-specific, mutable and lasting recognition of diverse foreign molecules. In their membrane-bound form, they are referred to as B cell receptors (BCRs): the collection of antibodies or BCRs in an individual is a record of their antigenic history, and tells us how their B cells have recognised and interacted with pathogens over their lifespan. Since the advent of next-generation sequencing and its application to these repertoires, we can sample and sequence the receptors of 103 to 106 B cells from a given individual. However, there remains an unmet need for mapping the mounting number of these BCR sequences to their putative antigen specificity. In this thesis, we aim to demonstrate how computational structural methods can be applied to BCR repertoires to improve our ability to identify antigen-specific antibodies.</p> <p>In the first chapter, we describe a novel method for identifying antigenspecific antibodies from repertoire data using paratope prediction which we call "paratyping", and then use paratyping to discover novel Pertussis toxoidbinding antibodies from the BCR repertoires of transgenic mice.</p> <p>Transgenic mice are a common source of therapeutic antibodies. To improve our understanding of the structural landscape of the BCR repertoires of these workhorses of antibody discovery, in the second chapter we apply structural annotation and modelling approaches to naive repertoires from humans, mice and transgenic mice. We show that the starting structural repertoires of transgenic mice are intermediate between humans and mice, despite being encoded by human genes, as a result of deficient junctional diversification in the primary repertoire of transgenic mice.</p> <p>In the third chapter, we describe the creation of a novel sequence database for Ebolavirus-binding antibodies called EBOV-AbDab. Using this data alongside other techniques, we analysed the longitudinal BCR repertoires of 40 individuals vaccinated with Ad26.ZEBOV/MVA-BN-Filo. We used paratyping to predict the epitopes of the majority of the most expanded clonotypes after the booster vaccine, providing evidence of a highly convergent response to Ebolavirus vaccination that correlated with anti-Ebolavirus glycoprotein IgG titre.</p> <p>Finally, in the fourth chapter we examine the growing body of nativelypaired data to try to understand VH:VL pairing preferences. We re-examine the VH:VL interface in the context of germline gene usage and subsequently analyse at scale the genetic pairing preference of antibodies in the largest paired sequencing repository, the Observed Antibody Space, in combination with 3x as much novel data produced by an alternative single-cell sequencing technology.</p> <p>In the final chapter, I outline further work which can extend or complete the work presented in the previous chapters, as well as general outlooks for the field.</p> |
| spellingShingle | Richardson, E Computational methods for rapid structural modelling of antibody-antigen interactions to improve identification of antigen-specific antibodies from BCR-seq data |
| title | Computational methods for rapid structural modelling of antibody-antigen interactions to improve identification of antigen-specific antibodies from BCR-seq data |
| title_full | Computational methods for rapid structural modelling of antibody-antigen interactions to improve identification of antigen-specific antibodies from BCR-seq data |
| title_fullStr | Computational methods for rapid structural modelling of antibody-antigen interactions to improve identification of antigen-specific antibodies from BCR-seq data |
| title_full_unstemmed | Computational methods for rapid structural modelling of antibody-antigen interactions to improve identification of antigen-specific antibodies from BCR-seq data |
| title_short | Computational methods for rapid structural modelling of antibody-antigen interactions to improve identification of antigen-specific antibodies from BCR-seq data |
| title_sort | computational methods for rapid structural modelling of antibody antigen interactions to improve identification of antigen specific antibodies from bcr seq data |
| work_keys_str_mv | AT richardsone computationalmethodsforrapidstructuralmodellingofantibodyantigeninteractionstoimproveidentificationofantigenspecificantibodiesfrombcrseqdata |