“AI tool predicts virus mutations, aids vaccine design during pandemics.”
An artificial intelligence tool has been created for the purpose of anticipating the emergence of new virus variants before they materialize. This remarkable development, a collaborative effort between the University of Oxford and Harvard Medical School, claims the capability to have foreseen mutations in the Covid-19 virus during the ongoing pandemic. Named EVEscape, this model holds the potential to contribute significantly to vaccine design by analyzing how viruses adapt in response to the human immune system.
The University of Oxford underscores the tool’s capacity to “predict the future.” Throughout the course of the Covid-19 pandemic, various virus variants emerged, characterized by multiple genetic alterations. These mutations can significantly influence the virus’s behavior, potentially enabling it to spread more rapidly or evade our immune system’s recognition and defense mechanisms. Notably, in late 2021, the Omicron variant exhibited such traits, infecting millions, albeit without causing a substantial surge in hospitalizations and fatalities.
EVEscape, which stands for Evolutionary Model of Variant Effect, combines a deep learning model of virus evolution with detailed biological and structural data. As outlined in the journal Nature, the research team elucidates how the model functions by forecasting the probability of a viral mutation enabling it to evade immune responses, such as preventing antibodies from binding. Notably, the model was tested using data available at the outset of the Covid-19 pandemic in February 2020 and successfully predicted which SARS-CoV-2 mutations would manifest and which would become predominant.
Furthermore, the model also anticipated which antibody-based treatments would lose their effectiveness as the pandemic unfolded and the virus underwent mutations to evade these therapies. The potential of this technology lies in its capacity to aid in preventive measures and the formulation of vaccines targeting concerning variants before they become prevalent.
Pascal Notin, co-lead author of the study, asserted that had this tool been deployed at the inception of the pandemic, it would have accurately foreseen the most prevalent Covid-19 mutations. He emphasized the significance of this work for pandemic surveillance and vaccine design, particularly in the context of emerging high-risk mutations.
