Machine learning-enhanced immunopeptidomics applied to T-cell epitope discovery for COVID-19 vaccines

Pillar 4
Functional Genomics & Structure-Function of VOCs

Next-generation T-cell-directed vaccines for COVID-19 are focusing on improving long-lasting immune protection by targeting T-cells, which play a key role in fighting infections. One of the challenges in designing these vaccines is identifying which parts of the virus that are recognized by T-cells (called T-cell epitopes) are most likely to trigger a strong immune response. This study introduces a new computational method to help pinpoint these critical T-cell epitopes more accurately.

The researchers developed a tool called MHCvalidator, which uses machine learning to enhance the ability to identify T-cell epitopes that recognize viral proteins. This tool was tested on a large dataset of JY cells (a line of immortalized white blood cells) and COVID-19 patient blood samples, showing that certain viral proteins, including a unique part of the spike protein, are present in small but notable proportions of cases. This discovery highlighted that some people’s immune systems are exposed to “frameshifted” versions of the spike protein, which could be important for future vaccine designs.

The study also introduced a tool called EpiTrack, which tracks changes in the viral T-cell epitopes over time as new variants of the virus, like Delta and Omicron, emerge. Most of the T-cell epitopes identified by the vaccine remain stable and recognizable by the immune system, but the researchers found that one particularly important epitope (linked to the HLA-A1 gene) mutates in the newer variants. This suggests that, while many parts of the virus stay the same, some mutations are evolving in a way that could impact how well the immune system recognizes and responds to the virus.

The findings emphasize the need for ongoing research to adapt vaccines to these evolutionary viral changes. This study provides valuable insights for the development of T-cell-based vaccines, which could offer better protection against not only current variants but also future ones, by focusing on parts of the virus that are less likely to evolve.

Read article