AI can help predict mutations of viruses before they occur

Viruses like SARS-CoV-2, which causes COVID-19 disease, mutate over time, which make some treatments and prevention measures ineffective. Those mutations are basically evolution happening right in front of our eyes and they are difficult to predict. Now scientists at the University of Waterloo  have developed an artificial intelligence-based method to foresee the most likely mutations of pathogens.

Artificial intelligence can analyse genes of the virus and predict the most likely mutations. Image credit: NIH Image Gallery via Wikimedia

If you can prepare for the threat before it emerges, you have better chances to defend yourself. If we knew what kind of pathogens are going to be bothering humanity in the very near future, we could have vaccines and treatments ready for the wave. Sadly, however, in most cases we can only address new variants and mutations after they emerge. This hinders our vaccine, treatment and test development. But now scientists focused on using a single ancestral sequence, which doesn’t change significantly when mutations occur, and collected all the surrounding information to train a neural network to predict the most likely mutations of the genome.

Artificial intelligence (AI) algorithm basically anticipates what mutations are most likely to occur in a given setting. It finds the path genome of the pathogen is going to take to adapt to the changing environment. Retrospectively, AI predicted and identified the variants that came to be known as alpha, beta, gamma, delta and other variants of concern. Obviously. this information to us is not that valuable, but at the beginning of the pandemic it would have been unbelievably useful. However, now we can use this tool for predicting mutations of other pathogens. In fact, scientists have much bigger ambitions for this AI-based approach.

Michelle Przedborski, another of the team members and a professor of applied mathematics at Waterloo, said: “Even with cancer, we should be able to identify the therapeutic targets for overcoming mutation-driven drug resistance. Lots of drugs are targeting a specific part of the protein in cancer cells. But if there are mutations in those, then drugs wouldn’t be effective anymore. We can apply the same analysis and AI method to other pathogens”.

The name of the game is being prepared. COVID-19 caused us a lot of trouble not because it is a particularly deadly disease, but because we were not prepared for it. People don’t have any supplies or a plan, healthcare institutions are not geared towards unexpected workloads, economies are not building with a hard future in mind. AI can help us prepare for these shocks at least in terms of vaccines, treatments and forecasts. 


Source: University of Waterloo