KIMPESE, a mid-sized town in southwestern Democratic Republic of the Congo, rings with a cheerful melody as the thumping bass of café sound systems mixes with the chatter from vendors selling bright fruits and soft fabrics. Against the background of this everyday chorus, it’s almost impossible to hear the tall man with grey-tinged stubble, standing beneath the overhang of a single-story, freshly painted building, talking to the bats overhead.

He speaks in a hybrid language of chirps and the local Kikongo dialect, cooing toward the outdoor rafters as mothers steer their children around the eccentric visitor and into the facility — the local health center — stealing curious but wary glances. The animals respond with a flurry of clicks and squeaks. “They’re happy to see us,” he says, laughing softly.

The irony of potentially disease-carrying bats lingering in the rafters of the local health center is not lost on Mulembakani.

Visual by Jeffrey Marlow/Undark

The bat-whisperer is Dr. Prime Mulembakani, and along with a couple of colleagues from the DRC’s Institut National de Recherche Biomédicale (INRB) in Kinshasa, he is in Kimpese to test bats for potentially dangerous viruses as part of a U.S. Agency for International Development effort called Predict. Launched in 2009 as part of USAID’s Emerging Pandemic Threats program, the Predict project was renewed in 2014 with a grant of $100 million over five years to develop a sprawling, searchable database of the zoonotic pathogens behind emerging pandemic threats in countries around the world.

If scientists can detail the places where lethal viruses simmer in wait, the thinking goes, they can head off a swelling pandemic and better manage outbreaks while they are still small and local. Researchers and other outbreak responders could consult this database to begin mapping the source of an emerging disease, for example, and quickly get to work on minimizing transmission and developing potential new vaccines that could save countless lives.

“We have a job to do,” Mulembakani explains. “We also have the opportunity to be in contact, in close contact, with people who are on the front line — the communities who are really at risk for a virus spillover from animals into people.” The irony of potentially disease-carrying bats hanging from the rafters of the local health center is not lost on Mulembakani, an epidemiologist by training, and he pauses for emphasis: “We need to stop these events from getting out of control.”

It’s an ambitious idea, and the stakes are high. The infamous 2014 Ebola outbreak, after all, likely began when a toddler came in contact with bat droppings while playing near a tree in rural Guinea. The virus went on to kill more than 11,000 people over the next two years, orphaning as many as 30,000 children and damaging affected West African economies. Similarly, the 2009 H1N1 “swine flu” pandemic that caused the deaths of nearly a quarter million people may have started when a five-year-old became infected in a mountain village 120 miles east of Mexico City. HIV/AIDS — the most devastating spillover of the last century — is credited with an estimated 35 million victims after the virus jumped from primates to humans in the 1920s.

And just two weeks ago, the World Health Organization declared another Ebola outbreak — this time in the northern part of the DRC.

Experts fear that spillover events like these are going to become more common in coming years. The increasing rate of human encroachment into animal habitats is removing the old buffers, making contact with potentially infected animals more common. And the math is simple: More points of contact facilitate more spillovers, and more transmission of animal pathogens into the human population. Add to that the globalized transportation network, in which billions of people (or potential hosts, from a virus’s perspective) are just a plane ride away, and the risks appear to grow at an incendiary rate.

That may sound daunting, but the era of “big data” virology is on the horizon, and it is now helping to drive embryonic efforts like the one here in Kimpese. “Right now, for example, we try to separately develop vaccines for SARS and MERS, which are both coronaviruses,” says Eddy Rubin, chief science officer at Metabiota, a San Francisco company that oversees several of Predict’s operations, including the one in the DRC. “If we had 10,000 coronavirus sequences, it’s likely we would identify common features, so that instead of making vaccines against individual viral species, we could be making vaccines against full viral families.”