Researchers say artificial intelligence could help predict the next virus to jump from animals such as bats, like the ones pictured, to humans. Photo by Daniel Streicker, Mollentze N, et al./PLOS Biology
Sept. 28 (UPI) -- Scientists are constantly monitoring the threat of zoonotic diseases, but there are millions of viruses circulating among animal populations, making the task especially challenging.
To increase the odds that researchers can identify the next virus to jump from animals to humans -- before it can spark a global pandemic -- scientists are enlisting the help of sophisticated algorithms and machine learning.
In a new proof-of-concept study, published Tuesday in the journal PLOS Biology, researchers suggest artificial intelligence can be used to predict the likelihood that an animal-infecting virus will infect humans.
To build the new machine-learning model, researchers first compiled a database of 861 zoonotic virus species from 36 families. Scientists then use artificial intelligence to train their model to identify genomic patterns linked with risk of human infection.
To test the new model's efficacy, scientists used it to analyze the dangers posed by another group of virus species not included in the original dataset.
"Our model reduced a second set of 645 animal-associated viruses that were excluded from training to 272 high and 41 very high-risk candidate zoonoses and showed significantly elevated predicted zoonotic risk in viruses from nonhuman primates, but not other mammalian or avian host groups," researchers wrote in the paper.
The modeling experiment showed genomic patterns are more predictive of the potential for human infection than a virus species' taxonomic relationships.
The model even successfully identified the virus that causes COVID-19, SARS-CoV-2, as a "relatively high-risk coronavirus" without any prior knowledge of other SARS-related coronaviruses.
While animal populations host millions of viruses, studies suggest only a small percentage hold the potential for human infection.
Machine-learning can make it easier for scientists to rule out non-threatening viruses and home in on those posing the greatest threat.
Once potential threats are identified by artificial intelligence models, scientists can use other investigation techniques to more precisely anticipate the threat and develop mitigation strategies.
"These findings add a crucial piece to the already surprising amount of information that we can extract from the genetic sequence of viruses using AI techniques," study co-author Simon Babayan said in a press release.
"A genomic sequence is typically the first, and often only, information we have on newly-discovered viruses, and the more information we can extract from it, the sooner we might identify the virus' origins and the zoonotic risk it may pose," said Babayan, a researcher at the University of Glasgow.
"As more viruses are characterized, the more effective our machine learning models will become at identifying the rare viruses that ought to be closely monitored and prioritized for preemptive vaccine development," Babayan said.