Viral Forecasting with AI

Researchers have developed a deep learning framework to forecast viral escape and put it to the test in a new study (1). Specifically, the team wanted to know how well the model – named “EVEscape” –  could predict SARS-CoV-2 viral escape with only the information available to us in February 2020, when our understanding of the virus was still in its infancy.

The model demonstrated high accuracy at predicting viral escape mutations that occurred during the pandemic, and even matched the accuracy of high-throughput experimental scans. In the paper, the researchers highlight that EVEscape can be generalized to other viruses, such as influenza and HIV, and to viruses with pandemic potential that have been largely overlooked in research, such as Lassa and Nipah virus.

“The critical aspect that makes our approach very powerful compared to traditional methods is that all the information we use in the EVEscape model is available at the very beginning of a pandemic,” said Yarin Gal, Associate Professor at the University of Oxford and a contributing author on the study, in a press release (2). “We developed new AI methods that do not have to wait for relevant antibodies to arise in the population to predict which variants are the most concerning.”

High genetic diversity from viral mutation and recombination makes developing vaccines and treatments during a pandemic particularly difficult, but having the ability to predict potential mutations could support the rapid development of more effective therapeutics. “Antibody escape mutations affect viral reinfection rates and the duration of vaccine efficacy,” said Gal (2). “Therefore, anticipating viral variants that avoid immune detection with sufficient lead time is key to developing optimal vaccines and therapeutics – and this is what EVEscape would enable us to do.”