Immunoinformatic identification of CD8+ T-cell epitopes
<p>Antigen-specific T-cells play a crucial role in the adaptive immune response by providing a defence mechanism against pathogens and maintaining tolerance against self-antigens. This sparked interest in the development of epitope-based vaccines and immunotherapies that elicit antigen-specifi...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2022
|
| Subjects: |
| _version_ | 1797111767639261184 |
|---|---|
| author | Lee, CH-J |
| author2 | Koohy, H |
| author_facet | Koohy, H Lee, CH-J |
| author_sort | Lee, CH-J |
| collection | OXFORD |
| description | <p>Antigen-specific T-cells play a crucial role in the adaptive immune response by providing a defence mechanism against pathogens and maintaining tolerance against self-antigens. This sparked interest in the development of epitope-based vaccines and immunotherapies that elicit antigen-specific T-cell responses. However, screening the antigens driving the response is currently labour-intensive, low-throughput and costly. Due to the limitations of experimental approaches, computational methods for predicting CD8+ T-cells have started to emerge. However, predicting the T-cell recognition potential of MHC-presented peptides has shown to be more challenging than predicting MHC ligands, and the full spectrum of features underlying peptide immunogenicity remains to be explored. Hence, this thesis presents a systems biology approach to study features of peptide immunogenicity and accurately predict CD8+ T-cell epitopes from HLA-I presented pathogenic or cancer peptides.</p>
<p>The thesis begins with an immunoinformatic analysis of antigen-specific T-cell profiles in the contexts of autoinflammatory and infectious diseases. In autoinflammatory disease, the multi-modal single-cell sequencing of ulcerative colitis and checkpoint treatment-induced colitis revealed pathology-specific differential expressions of cytotoxic T-cells. The current technologies, however, were unable to identify the source antigen, emphasising the importance of predicting T-cell targets to better understand disease pathology. Moreover, in infectious diseases, CD8+ T-cell epitope prediction algorithms facilitated the understanding of disease heterogeneity and vaccine design during the COVID-19 pandemic, but many existing algorithms were found to be ill-suited for predicting epitopes from emerging pathogens.</p>
<p>Therefore, a novel computational workflow was developed for an accurate and robust prediction of source antigens driving the cellular immune response. First, an unbiased evaluation of state-of-the-art algorithms revealed that they perform poorly on both cancer neoepitopes (e.g. glioblastoma) and pathogenic (e.g. SARS-CoV-2) epitopes. After investigating the reasons for low performance, TRAP, a deep learning workflow for context-specific prediction of CD8+ T-cell epitopes, was developed to effectively capture T-cell recognition motifs. The application of TRAP was demonstrated by using it to investigate the immune escape potential of all theoretical SARS-CoV-2 mutants. Thus, this thesis presents a novel computational platform for accurately predicting CD8+ T-cell epitopes to foster a better understanding of TCR:pMHC interaction and the development of effective clinical therapeutics.</p> |
| first_indexed | 2024-03-07T08:15:01Z |
| format | Thesis |
| id | oxford-uuid:013c2bdf-0455-48e0-a6c7-0768ec3fb80c |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T08:15:01Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | oxford-uuid:013c2bdf-0455-48e0-a6c7-0768ec3fb80c2024-01-04T08:45:16ZImmunoinformatic identification of CD8+ T-cell epitopes Thesishttp://purl.org/coar/resource_type/c_db06uuid:013c2bdf-0455-48e0-a6c7-0768ec3fb80cComputational immunologyEnglishHyrax Deposit2022Lee, CH-JKoohy, HSimmons, A<p>Antigen-specific T-cells play a crucial role in the adaptive immune response by providing a defence mechanism against pathogens and maintaining tolerance against self-antigens. This sparked interest in the development of epitope-based vaccines and immunotherapies that elicit antigen-specific T-cell responses. However, screening the antigens driving the response is currently labour-intensive, low-throughput and costly. Due to the limitations of experimental approaches, computational methods for predicting CD8+ T-cells have started to emerge. However, predicting the T-cell recognition potential of MHC-presented peptides has shown to be more challenging than predicting MHC ligands, and the full spectrum of features underlying peptide immunogenicity remains to be explored. Hence, this thesis presents a systems biology approach to study features of peptide immunogenicity and accurately predict CD8+ T-cell epitopes from HLA-I presented pathogenic or cancer peptides.</p> <p>The thesis begins with an immunoinformatic analysis of antigen-specific T-cell profiles in the contexts of autoinflammatory and infectious diseases. In autoinflammatory disease, the multi-modal single-cell sequencing of ulcerative colitis and checkpoint treatment-induced colitis revealed pathology-specific differential expressions of cytotoxic T-cells. The current technologies, however, were unable to identify the source antigen, emphasising the importance of predicting T-cell targets to better understand disease pathology. Moreover, in infectious diseases, CD8+ T-cell epitope prediction algorithms facilitated the understanding of disease heterogeneity and vaccine design during the COVID-19 pandemic, but many existing algorithms were found to be ill-suited for predicting epitopes from emerging pathogens.</p> <p>Therefore, a novel computational workflow was developed for an accurate and robust prediction of source antigens driving the cellular immune response. First, an unbiased evaluation of state-of-the-art algorithms revealed that they perform poorly on both cancer neoepitopes (e.g. glioblastoma) and pathogenic (e.g. SARS-CoV-2) epitopes. After investigating the reasons for low performance, TRAP, a deep learning workflow for context-specific prediction of CD8+ T-cell epitopes, was developed to effectively capture T-cell recognition motifs. The application of TRAP was demonstrated by using it to investigate the immune escape potential of all theoretical SARS-CoV-2 mutants. Thus, this thesis presents a novel computational platform for accurately predicting CD8+ T-cell epitopes to foster a better understanding of TCR:pMHC interaction and the development of effective clinical therapeutics.</p> |
| spellingShingle | Computational immunology Lee, CH-J Immunoinformatic identification of CD8+ T-cell epitopes |
| title | Immunoinformatic identification of CD8+ T-cell epitopes
|
| title_full | Immunoinformatic identification of CD8+ T-cell epitopes
|
| title_fullStr | Immunoinformatic identification of CD8+ T-cell epitopes
|
| title_full_unstemmed | Immunoinformatic identification of CD8+ T-cell epitopes
|
| title_short | Immunoinformatic identification of CD8+ T-cell epitopes
|
| title_sort | immunoinformatic identification of cd8 t cell epitopes |
| topic | Computational immunology |
| work_keys_str_mv | AT leechj immunoinformaticidentificationofcd8tcellepitopes |