A University of Liverpool study could help scientists mitigate the future spread of zoonotic and livestock diseases caused by existing viruses. The researchers applied a machine learning (ML) artificial intelligence (AI) technology to predict more than 20,000 unknown associations between known viruses and susceptible mammalian species. They suggest the findings, published in Nature Communications, could be used to help target disease surveillance programs.
“As viruses continue to move across the globe, our model provides a powerful way to assess potential hosts they have yet to encounter,” said Maya Wardeh, PhD, from the University’s Institute of Infection, Veterinary and Ecological Sciences. “Having this foresight could help to identify and mitigate zoonotic and animal-disease risks, such as spill-over from animal reservoirs into human populations.” Wardeh and colleagues report on their study in a paper titled, “Divide-and-conquer: machine-learning integrates mammalian and viral traits with network features to predict virus-mammal associations.”
Thousands of viruses are known to affect mammals, with recent estimates indicating that less than 1% of mammalian viral diversity has been discovered to date, the authors reported. Some of these viruses such as human and feline immunodeficiency viruses have a very narrow host range, whereas others such as rabies and West Nile viruses have very wide host ranges. Rabies could theoretical infect any mammal, the team pointed out.
“Host range is an important predictor of whether a virus is zoonotic and therefore poses a risk to humans,” Wardeh continued. “Most recently, SARS-CoV-2 has been found to have a relatively broad host range which may have facilitated its spill-over to humans. However, our knowledge of the host range of most viruses remains limited.”
And what information we do have is “hugely biased towards humans and domesticated animals,” the authors noted. To address the knowledge gap, the researchers developed a novel machine learning framework to predict unknown associations between known viruses and susceptible mammalian species. “Our framework to predict unknown associations between known viruses and potential mammalian hosts or susceptible species comprised three distinct perspectives: viral, mammalian and network,” the team noted. “Each perspective produced predictions from a unique vantage point (that of each virus, each mammal, and the network connecting them respectively).
Their ML approach predicted over 20,000 unknown associations between known viruses and mammalian hosts, “… suggesting that current knowledge greatly underestimates the number of associations between wild and semi-domesticated mammals.” The data indicated that there are more than five times as many associations between known zoonotic viruses and wild and semi-domesticated mammals than previously thought. The model also predicted a five-fold increase in associations between wild and semi-domesticated mammals and viruses of economically important domestic species such as livestock and pets.
“Overall, we predict a 5.35-fold increase in associations between wild and semi-domesticated mammalian hosts and known zoonotic viruses (found in humans, excluding rabies virus),” they wrote. “Similarly, our results indicate a 5.20-fold increase between wild and semi-domesticated mammals and viruses of economically important domestic species (e.g. livestock and pets).
In particular, bats and rodents, which have been associated with recent outbreaks of emerging viruses such as coronaviruses and hantaviruses, were linked with increased risk of zoonotic viruses. Wild ruminants were also a key consideration. “Analogous to bats and rodents being important hosts of zoonotic viruses, wild ruminants are key in the maintenance and circulation of viruses affecting ruminant livestock,” the team noted. “Our framework highlights this knowledge-gap by predicting a 7.77-fold increase in number of associations between wild and semi-domesticated ruminants and known viruses.”
Commenting on the study, Wardeh said, “As viruses continue to move across the globe, our model provides a powerful way to assess potential hosts they have yet to encounter. Having this foresight could help to identify and mitigate zoonotic and animal-disease risks, such as spill-over from animal reservoirs into human populations.” The authors concluded, ““We applied a divide-and-conquer approach which separated viral, mammalian and network features into three unique perspectives, each predicting associations independently to enhance predictive power … Completing the picture of virus-host interactions can help identify and mitigate current and future zoonotic and animal-disease risks, including spill-over from animals into humans.”
Wardeh is currently expanding the approach to predict the ability of ticks and insects to transmit viruses to birds and mammals, which will enable prioritization of laboratory-based vector-competence studies worldwide to help mitigate future outbreaks of vector-borne diseases.
