myNEO has evaluated whether symptom severity is related to immunogenicity across three different types of Corona viruses
Many of us are wondering where we are in finding a treatment for the recent Covid-19 (SARS-CoV-2) outbreak that is spreading all over the world. The papers and social media are overloaded with press releases and articles, making it difficult to have an overview of who is doing what in this race against Covid-19. This has inspired us to give you an overview of what is happening in the field right now.
In addition, intrigued by the Covid-19 virus, the team of myNEO decided to conduct a small case study to test the effectiveness of its already developed (oncology) tools in a more broad immunology setting. More specifically, we’ve evaluated the hypothesis whether a more immunogenic virus indeed results in higher symptom severity, compared across 3 different types of Corona viruses. In parallel, the myNEO immunogenicity prediction algorithms have been validated by distinguishing immunogenic from non-immunogenic viral proteins.
The myNEO ImmunoEngine pipeline is designed to accurately detect high-confidence tumor-derived neoepitopes of immunotherapeutic purposes. To that end, myNEO has developed machine learning-based tools that allow to predict whether a specific peptide is likely to be presented at the cell surface, as well as whether such peptides are likely to elicit an immune reaction. Routinely, these tools are used to determine which tumor-specific peptides are the best candidates for vaccine development. However, it was deemed interesting to see whether the trained algorithms could also be used to gain further insight on the COVID-19 disease, caused by the coronavirus SARS-CoV-2. The virus was compared to SARS-CoV and MERS-CoV, the two other viruses from the coronavirus family that caused the SARS and MERS outbreaks in the last decades.
These three viruses are closely related, and yet elicit symptoms of varying orders of magnitude – symptoms of SARS-CoV were more severe than for the current coronavirus, while MERS-CoV was even more deadly than both of them. It is thought (as mentioned in McManus et al., 2014) that high immunogenicity could positively correlate with symptom severity, so it was investigated whether there was any difference in immunogenicity in relevant virus-derived peptides from each of these viruses.Figure I: Histogram and associated density curves of immunogenicity probability for the envelope-derived cell-surface-presented peptides of SARS-CoV-2 (red), SARS-CoV (blue) and MERS-CoV (green).
After gathering the sequences of the envelope proteins of SARS-CoV-2 as well as for SARS-CoV and MERS-CoV, all possible 9-mers were extracted. These were run through our presentation prediction algorithm to determine which ones are likely to be presented at the surface of infected cells. Subsequently, our proprietary neoIM algorithm ranked the presented peptides according to their predicted immunogenicity. This reflects the population-wide likelihood of this peptide to elicit an immune reaction. The resulting immunogenicity likelihood distribution of this reduced set of presented peptides was examined, and as can be seen in the graph above, SARS-CoV-2 exhibits on average fewer potentially immunogenic peptides than the SARS-CoV, which in turns has fewer likely immunogenic peptides than MERS-CoV. This implies that the immunogenic load of these three peptides does seem to correlate with their pathogenicity!
Although limited in scope, this quick case study constituted an interesting use of our proprietary tools, showcasing their adaptability and situation-wise robustness.