We didn't know about the SARS-CoV-2 virus until it showed up in humans. But previous experience with other coronaviruses that jumped into humans (SARS and MERS) told us that something like COVID-19 could pose a risk. Coronaviruses are prevalent in a number of species that have frequent contact with humans, and they have a clear history of being able to adapt themselves to human cells.
Being aware of viruses that have similar properties can help us recognize threats for future pandemics. Now, researchers are taking the results of a massive virus survey and releasing a public database of hundreds of viruses, all rated for how much risk the viruses pose to humans. And any viruses that we discover can be plugged into the framework the researchers have developed so that we can get quick information on whether they're threatening.
What’s out there?
The effort grew out of a USAID-sponsored program called PREDICT, which was part of a set of efforts focused on zoonotic diseases, those that can cross species barriers and infect humans. Collectively, the PREDICT project did a massive survey of animal viruses, using over a half-million individual samples taken from 75,000 animals. Out of that data, the project identified over 700 new viruses, including one that had never been seen in the animal in which it was found.
Knowing the genome sequence of the viruses doesn't tell us much about the risk the viruses pose to humans on their own. We can figure out what proteins the viruses encode, but we're not able to look at those proteins and figure out whether they make the virus any more likely to infect humans. And it's not only infectivity that poses a risk. If a virus normally circulates in rare animals that avoid humans, the chances of it jumping to us is small.
Factors abound, as does disagreement among experts about how important those factors are. So figuring out how to evaluate the new viruses posed a challenge.
To figure out what's important, the researchers got 150 virology and public health experts to consider 50 different potential risk factors, including the host species carrying a virus, the location in which species was found, and the species' evolutionary relationship to known viruses. The experts were asked to rank the importance of each of these risk factors, and the PREDICT team weighed each of its ratings based on the person's expertise in each issue. (For example, the opinion of a virologist might count less on issues related to how often its host animal interacts with humans.)
Some of the important risk factors that were consistently rated highly were obvious: frequency of interactions with humans and our livestock, the ability to infect a variety of hosts, and modes of transmission. Not every factor was rated as very significant, and seven were rated as important. But we simply don't have enough data on most viruses to evaluate them properly.
Scoring spillover
The net result is a spillover score, the best estimate of the risk each of these viruses pose to humans, awkwardly rated on a score of 1 to 155 (this is what happens when you start with 50 factors scored from 1-5, weigh them to varying degrees, and then throw some of them out). As a test of the system's validity, the researchers looked at the top-scoring viruses; all of the first dozen were already known to have infected humans.
SARS-CoV-2 ranked right between two viruses that have caused multiple outbreaks of hemorrhagic fever in Africa: Lassa and Ebola. It didn't come out on top because those other viruses have caused multiple outbreaks (SARS-CoV-2 has only had one, but it has made it count). We also know a lot more about other viruses' normal hosts, while we haven't identified the host species of SARS-CoV-2 before it moved into humans.
All of the analysis is being made available via the Spillover website, which includes a ranked list of all the viruses analyzed so far. A quick view of each breaks out the risk into three categories (based on the host it's found in, the environment of that host, and the virus's genetics). A detailed view breaks out each individual factor that we have enough data to evaluate.
Beyond making the data on these new viruses available, Spillover is also a flexible sharing platform. "Flexible" in that, as we learn more about what makes a virus a zoonotic threat, the researchers promise they'll update the analyses for all the viruses in the database. And "sharing" because the PREDICT team hopes that the research community will add new viruses to be rated as they're discovered. It's possible to create a risk score with as few as half a dozen viral properties.
While tracking this many new viruses is a great start, there are a few obvious limitations. For one, because they're already tracked intensively, influenza viruses are not in the database. Second, while it represents a lot of work, the hundreds of viruses described here are a drop in the bucket compared to the estimated 1.7 million viruses that infect mammals and birds. We've got a lot more work to do if we really want to avoid having the next pandemic sneak up on us.
Still, the project represents a valuable start. Several of the viruses that hadn't been described before are rated as more threatening than viruses that we already know can make the leap into humans. Obviously, targeting those viruses for study and more careful surveillance has the potential for a significant payoff, especially when compared to the global costs of the COVID-19 pandemic.
PNAS, 2021. DOI: 10.1073/pnas.2002324118 (About DOIs).
Listing image by STEEVE JORDAN / AFP / Getty Images
reader comments
29