Featured image
Healthcare

Predicting Virus Mutations: New AI Tool Developed by Harvard and Oxford Researchers

avatar

Sven

October 15th, 2023

~ 4 min read

In the ongoing battle against the COVID-19 pandemic, scientists have been racing to stay one step ahead of the ever-evolving SARS-CoV-2 virus. Now, a team of researchers from Harvard Medical School and the University of Oxford has developed a groundbreaking artificial intelligence (AI) tool that can predict viral variants before they emerge. Named EVEscape, this powerful tool combines a model of evolutionary sequences with detailed biological and structural information about the virus to make accurate predictions about the most likely variants to occur.

Published in the journal Nature, the study showcases how EVEscape would have successfully predicted the most frequent mutations and identified the most concerning variants of SARS-CoV-2 if it had been deployed at the beginning of the pandemic. Notably, EVEscape has also demonstrated its ability to accurately predict variants of other viruses such as HIV and influenza. This revolutionary tool provides valuable insights into the future of viral evolution and could aid in the development of more effective vaccines and therapies.

The foundation of EVEscape lies in a previous tool called EVE, short for evolutionary model of variant effect, which was originally developed to detect disease-causing gene mutations. By leveraging large-scale evolutionary data, EVE could predict the functionality of proteins based on genetic variations across species. Building upon this success, the researchers behind EVEscape saw an opportunity to repurpose the tool for predicting viral variants.

To develop EVEscape, the researchers enhanced the generative model from EVE by incorporating biological and structural details about the virus, including information about regions most easily targeted by the immune system. This comprehensive approach allows EVEscape to be flexible and adaptable to any virus, making it a powerful tool for predicting future variants.

In a fascinating experiment, the researchers turned back the clock to January 2020, just before the COVID-19 pandemic began, and tasked EVEscape with predicting the future of SARS-CoV-2. The results were impressive, with EVEscape accurately predicting which mutations would occur during the pandemic, outperforming experimental approaches that test the virus's ability to bind to antibodies made by the immune system.

EVEscape's predictions also extended to identifying which antibody-based therapies would lose their efficacy as the virus mutated to escape these treatments. Additionally, the tool demonstrated its capability to analyze the vast number of new SARS-CoV-2 variants and identify those most likely to become problematic. By rapidly determining the threat level of new variants, EVEscape can help inform earlier public health decisions and guide vaccine and therapeutic development.

The researchers are now applying EVEscape to real-time monitoring of SARS-CoV-2, using all available information to make predictions about its future evolution. They publish biweekly rankings of new SARS-CoV-2 variants on their website, sharing vital information with organizations like the World Health Organization. Moreover, the complete code for EVEscape is freely available online, promoting transparency and collaboration in the scientific community.

Beyond SARS-CoV-2, EVEscape has the potential to be generalized to other common viruses, including HIV and influenza. This versatility opens up exciting possibilities for predicting future viral variants and designing mutation-proof vaccines and therapies. By staying ahead of viral evolution, scientists can develop treatments that can withstand the escape mechanisms employed by viruses.

The development of EVEscape represents a significant breakthrough in our ability to predict viral variants and anticipate their impact on public health. With its accurate predictions and efficient capabilities, EVEscape provides a valuable tool for researchers and public health officials alike. By leveraging AI and evolutionary data, scientists can gain invaluable insights into the future of viral evolution, ultimately leading to the development of more effective vaccines and therapies.

Harvard Blog Post