Where is the next zoonotic virus hiding?

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

This new prediction came about for a machine learning method that spent 35,000 hours on a Calcul Quebec computer to find out how thousands of mammals – the host – and many viruses communicate.

After revealing 80,000 new potential interactions between the host and the virus, the network was then linked to a model consisting of viral genomic sequences to re-evaluate the possibility of transmission of all viruses in the database to humans.

Results: Creating a list of viruses of animal origin that pose a risk of infecting humans without causing genotypes.

The discovery, funded by IVADO, conducted as part of the Institute for Data Valuation, and the Viral Emergency Research Initiative, is the result of an international collaboration led by Timothy Poyst, professor of biology at the University of Montreal. In his research, he is interested in calculating the risk of future epidemics.

Focusing on the “forgotten” virus

Timothy Poisson

Credit: Amelie Phillibert

To verify their predictions, Poist and his virology, AI and public health experts have examined the literature to see if the selected viruses have ever been prevalent in humans. And 11 of the 20 viruses with the highest zoonotic potential actually made people sick.

“Some viruses really surprised us, we didn’t think they could be transmitted to humans,” said Timothy Poissot. In the case of the virus murine ectromelia, which is responsible for smallpox in rats. Our system gave it a “very high” chance of infecting humans, and we found that a Chinese school had an outbreak of the virus in 1987, but it was not listed in any database.

In general, the most common families of viruses are Bunivirus (Rift Valley Fever), Rhabdovirus (Rabies), Philovirus (Ebola Fever) and Flavivirus (Dengue Fever, Yellow Fever). “These are all families that are recognized for their significant zoonotic risk, but the model may allow us to measure risk more precisely within this family,” notes Timothy Poissot.

Monitor hotspots with Amazon …

The goal of this effort to predict viral infections is to direct the work of virologists in zoological prevention, which can lead to epidemics or even pandemics. The list of monitored viruses allows, among other things, to guide them in their sample propagation, as it targets the species and also determines their geographical distribution, the research team maps the results.

“As an ecologist who does biogeography, it’s important to know not only which virus will be compatible with which host, but also where we can find these combinations,” said Timothy Poist.

According to computer system results, Amazon is one of the places in the world where viral evolution is most likely. “The results are unequivocal: this is a hotspot in terms of the originality of host-virus interactions, that is, the region has the most interactions that do not normally occur”, underlines the investigator.

He explains these special acquaintances by the lack of virus information on the Amazon, deforestation, climate change and the expansion of cities, which increase communication between animals and humans.

Leave a Comment