Q&A: New Tool Ranks Viruses by Their Risk of Jumping to Humans
Q&A: New Tool Ranks Viruses by Their Risk of Jumping to Humans

Q&A: New Tool Ranks Viruses by Their Risk of Jumping to Humans

Researchers involved in a long-term project to identify viruses of concern have now assessed the risk factors that could help forecast which pathogens are the most likely to spillover from wildlife to people.

Jef Akst
Jef Akst
Apr 9, 2021

ABOVE: © ISTOCK.COM, AEROGONDO

Well before the world began grappling with the COVID-19 pandemic, researchers were already looking out for potential outbreaks from emerging diseases—and trying to stop them. A major hurdle in doing so is understanding which viruses in animals are most likely to make the jump to people. A new, interactive web-based tool, published April 5 in PNAS, uses 32 risk factors and data on more than 500,000 samples taken from nearly 75,000 animals, along with public records of virus detections in wildlife, to rank the chances of spillover among 887 viruses.

Project leader Jonna Mazet, an epidemiologist and disease ecologist at the University of California, Davis’s School of Veterinary Medicine, spoke with The Scientist about the “SpillOver” tool that she and her collaborators developed.

The Scientist: Tell me about how this project got started.

Jonna Mazet
UC DaVIS

Jonna Mazet: For more than a decade, I have been the PI and leader of the PREDICT Consortium, which is a very large group of scientists and laboratorians and public health professionals working in more than 35 countries around the world to strengthen systems to identify viruses of concern before they spill over and make people sick. And in doing that work, we were strengthening the systems, but we were also discovering viruses, and we wanted to understand and put some information for the policymakers around the risk of the viruses we were finding.

I think we were a little surprised and disappointed to find that no good information was in the scientific literature about really how to rank these viruses. So we had to start that effort as we were building the systems and as we were discovering viruses. This is a culmination of that huge collaborative project that included at least 400 individuals in the PREDICT project as well as experts from around the world in virology, ecology, epidemiology, and other disciplines.

TS: How did you build the SpillOver tool, and how does it work?

JM: We did intensive literature reviews, and we also mined the minds, if you will, of the scientists and individuals working on the PREDICT project. And then we collated all of the risk factors we could identify as . . . bits of risk in all the scientific papers that have talked about viral spillover risk and even spread. . . . We added to those ones that we were finding in the PREDICT project, because for the most part, the ones we could find in the literature were only around virology and didn’t include the host, the environmental risk component for exposure, or any of the ecology. . . . And then we reached out to scientists all over the world that were working at the top of their fields in this specific area of zoonotic disease and virology and spillover, and we asked them to rank those risk factors that we had identified as well as rank their expertise.

So, for example, if a virologist was ranking one of the virology-oriented risk factors, they may rate themselves as an expert. But if they were looking at one that was more in the ecology realm, they might rate themselves a little lower in their expertise. And we use their rankings as well as their self-assigned expertise to then look at all of the risk factors and put together a program—equations, basically—to come up with a weighted score for each risk factor. And then we used that to then find the data for all of the known zoonotics that were found first in wildlife and transmitted to people as kind of a gut check of our ranking system to see if it was working. And then once we found that the tool looked to be functioning very well for historical spillovers, we then ranked the viruses that the PREDICT project found.

See “Predicting Future Zoonotic Disease Outbreaks

TS: Where did SARS-CoV-2 rank?

JM: When we were first working on this, obviously there was no SARS-CoV-2 that we knew of—it existed, but it hadn’t been identified yet. So initially, it wasn’t even in our system, but of course as we were coming to put the final touches on the manuscript and the tool, we did add SARS-CoV-2 . . . with all other viruses that were coming out in the literature and in GenBank and GISAID and others.

When we added SARS-CoV-2, it ranked number two of the known zoonotics—[second to Lassa virus, found among rodents in West Africa and which causes hemorrhagic fever in people]. That’s ranking for its ability and likelihood to spill over again, and it has a bit of a nod to pandemic potential with our risk-ranking system. And I think that’s very telling. . . . Obviously, it’s a terrible virus that’s caused the pandemic, so it should rank very highly, as it does. And the reason that it’s not ranking even higher as number one is that it hasn’t been studied, until it spilled over.

Our goal is to actually rank viruses and study them before they spill over, so that we have them ranked on a watch list, so that countries that have these viruses can create watch lists and do the surveillance and risk mitigation before they spill over. As more and more information is coming out about the host and the distribution of SARS-CoV-2—it’s obviously worldwide in people but we’re interested in its distribution in wildlife and the potential reservoir hosts—I think it might even go up to number one.

Z.L. Grange et al., “Ranking the risk of animal-to-human spillover for newly discovered viruses,” PNAS, 118:e2002324118, 2021.

Editor’s note: This interview has been edited for brevity.