A team of researchers from the University of Oxford and Harvard Medical School has developed an artificial intelligence (AI) tool that can help forecast new virus variants before they actually appear, using only the information available at the start of an outbreak. The findings were published in the journal Nature.
The AI tool, called EVEscape, works by estimating the likelihood of a viral mutation that can escape immune responses, such as by preventing antibodies from binding. EVEscape combines a deep learning model of evolutionary viral sequences with detailed biological and structural information about the virus, which enables it to make predictions.
This approach could facilitate more effective preventive action, as well as the design of vaccines that target variants of concern before they become prevalent.
“This work is of enormous value, both for pandemic surveillance efforts, but also for informing vaccine design in a way that is robust to the emergence of certain risk mutations. The most exciting next step for this line of work is demonstrating how it can be used in practice to inform vaccine design,” said study co-lead author Pascal Notin, a student in the Oxford Applied and Theoretical Machine Learning (OATML) group, part of the Department of Computer Science at the University of Oxford.
The researchers tested the ability of EVEscape to make advance predictions by inputting only information available at the start of the COVID-19 pandemic in February 2020. Based on genetic sequences for spike proteins from the Coronaviridae virus family, they asked EVEscape to predict what would happen to SARS-CoV-2.
The model successfully predicted which mutations occurred during the health crisis and which would become more prevalent. It also showed accuracy in predicting immune escape mutations for influenza, HIV, and some other trained pathogens with pandemic potential such as Lassa and Nipah.
“The critical aspect that makes our approach very powerful compared to traditional methods is that all of the information we use in EVEscape is available at the very beginning of a pandemic. We have developed new AI methods that do not need to wait for relevant antibodies to emerge in the population to predict which variants are of most concern,” explained contributing author Yarin Gal, Associate Professor at the University of Oxford.
The core component of EVEscape is EVE, short for “variant effect evolutionary model”, a deep generative model of protein sequences that helps researchers understand which mutations preserve the fitness of a given virus.
The research team originally developed EVE to predict the effects of genetic mutations on the risk of diseases such as cancer and heart disease. In previous work, it proved to be accurate.
“Unlike previous machine learning methods that were trained by learning from datasets manually annotated by clinicians, EVE learns in an unsupervised manner from a large collection of evolution-related protein sequences. This helps avoid the biases and limitations inherent in human annotation,” said senior author Debora Marks, professor of systems biology in the Blavatnik Institute at Harvard Medical School.
The researchers are now using EVEscape to look ahead at SARS-CoV-2 and predict future variants of concern; every two weeks, they release a ranking of new variants. Eventually, this information could help scientists develop more effective vaccines and therapies. The team is also broadening the work to include more viruses.
“We want to know if we can anticipate the variation in viruses and forecast new variants — because if we can, that’s going to be extremely important for designing vaccines and therapies,” Marks said.
With information from the University of Oxford, Image by Getting Images published on the University of Oxford website (Linked directly)
