Where are the following zoonotic viruses hiding?

Until recently, we only knew about two percent of the possible interactions between viruses and mammals, the so-called virome. However, a new artificial intelligence (AI) technique has revealed new possible viral interactions, increasing the size of the known virome by a factor of 15.

This new prediction came from a machine learning approach that spent 35,000 hours on Calcul Québec computers looking at information about how a thousand host mammals and so many viruses interact.

After identifying 80,000 new possible interactions between hosts and viruses, the network was then combined with a model made up of viral genomic sequences to re-evaluate the potential for infecting humans with all the viruses in the database.

Result: creation of a list of viruses of animal origin that pose a risk of human infection, except for zoonoses.

This discovery, funded by IVADO, the Data Valuation Institute, and carried out as part of the Virus Emergence Research Initiative, is the result of an international collaboration led by Timothée Poiseau, professor in the Department of Biological Sciences at the University of Montreal. In his research, he is interested in calculating the risks of a future pandemic.

Shed light on “forgotten” viruses

Timothy Poiseau

Credit: Amelie Philibert

To confirm their predictions, Poiseau and his team of experts in virology, artificial intelligence and public health checked the literature to see if the viruses that had been noted had ever been human outbreaks. And of the 20 viruses with the highest zoonotic potential, 11 actually disgusted people.

“Some viruses really surprised us, we didn’t think that they could be transmitted to people,” says Timothy Poiseau. This is a case of murine ectromelia, the virus responsible for smallpox in mice. Our system gave it a “very high” chance of infecting humans, and we found that there was indeed an outbreak of this virus in a Chinese school in 1987, but it was not listed in any database.

In general, the most prominent families of viruses include bunyaviruses (Rift Valley fever), rhabdoviruses (rabies), filoviruses (Ebola) and flaviviruses (dengue fever, yellow fever). “These families are all recognized for their significant zoonotic risk, but the model may allow us to more accurately measure the risk in these families,” notes Timothy Poiseau.

Access point monitoring including Amazon…

These viral infection prediction efforts aim to guide the work of virologists in the prevention of zoonoses that can lead to epidemics or even pandemics. The list of viruses to be monitored allows, among other things, to be sent to sampling campaigns, as it targets the species and also determines their geographic distribution, and the research team has mapped the results.

“As a biogeographical ecologist, it was important to know not only which virus would be compatible with which host, but also where we could find those combinations,” says Timothée Poiseau.

According to the results of the computer system, the Amazon is the place in the world where the greatest potential for viral evolution is. “The results are unequivocal: this is a hotspot in terms of the originality of virus-host interactions, that is, this region has the most interactions that usually do not occur,” the search engine emphasizes.

He attributes these special contacts to the lack of data on the virome in the Amazon, deforestation, climate change, and urban expansion, which increases contact between animals and humans.

Leave a Comment

Your email address will not be published.