Artificial intelligence could reduce the threat of future viral pandemics, according to the researchers who have developed a machine learning-based early warning system.
The scientists from the Scripps Research Institute have trained the system to track the emergence and evolution of epidemic viruses – like the SARS-CoV-2 variants.
This system could be used to track viral pandemics in the future, using an ‘unprecedented’ approach – according to the paper’s senior author Professor William Balch. “There are rules of pandemic virus evolution that we have not understood but can be discovered,” he said.
In a new paper published in Cell Patterns, the scientists show that this system could have predicted the emergence of new COVID variants weeks before the World Health Organisation (WHO) designated them as threats.
“One of the big lessons of this work is that it is important to take into account not just a few prominent variants, but also the tens of thousands of other undesignated variants, which we call the ‘variant dark matter’,” said Balch.
The AI was able to identify the key variants emerging from the ‘dark matter’, meaning those that significantly affected viral spread and mortality rates.
When applying the AI to data from the COVID pandemic, the machine was able to track genetic changes in the variants, as well as the virus’ adaptations to lockdowns, mask wearing, new vaccines, increasing human immunity, and competition between different variants.
The scientists hope their findings show that similar early warning systems could track the evolution of future viral pandemics in real-time. This could help scientists to predict increases in infection rates in time to prepare countermeasures, such mask wearing and healthcare service provisions.
The AI could also help in the race to find treatments and vaccines during pandemics, as the system also identified key COVID proteins and their roles in the evolution of the pandemic.
“This system and its underlying technical methods have many possible future applications,” said Dr Ben Calverley, co-first author of the study.
Read more: