Researchers at Boston-based Harvard Medical School and University of Oxford in England have created an AI tool to forecast which COVID-19 strains will grow in dominance, according to an Oct. 11 article in Nature.
The tool, called EVEscape, predicts how the virus can evolve through a model of evolutionary sequences alongside biological and structural data, according to an Oct. 11 Harvard news release. EVEscape works to forecast which future COVID-19 strains are most likely to occur.
Every two weeks, the researchers will release a ranking of COVID-19 variants.
When the COVID-19 public health emergency ended in May, national trackers of the pandemic shifted to reporting on hospital admission levels and COVID-19 deaths among all deaths, and the CDC stopped covering COVID-19 case counts and test positivity rates. In the months since, it has been more difficult to get a clear picture on how SARS-CoV-2 is mutating. The new AI tool could fill that deficit.
EVE, short for "evolutionary model of variant effect," was first made to detect gene mutations that cause human diseases. The researchers added biological and structural information about SARS-CoV-2 to create EVEscape. The first tool can thus be translated to other viruses, including influenza and HIV.
"As the framework is generalizable across viruses, EVEscape can be used from the start for future pandemics, as well as to better understand and prepare for emerging pathogens," the researchers said in Nature.
They found that if the tool launched at the start of the COVID-19 pandemic, it could have identified the most concerning strains before they emerged.
The rankings are available here.