Shazam for Mosquitoes

An unexpected call led one scientist to develop an easy way of tracking the animal that endangers more humans than any other.

Aedes aegypti  (Paulo Whitaker)

It was late on a Friday night, and Haripriya Mukundarajan was trying to record the annoying buzz of a mosquito’s wings.

To fight mosquitoes and the diseases they carry, you need to know where they are—which species, in which places. Everything else flows from that. A lot of labor goes into trapping, counting, and identifying the insects, but Mukundarajan figured there might be an easier way. Mosquitoes give themselves away through the vexing whine of every wingbeat. If Mukundarajan could find an easy way of tracking that sound, perhaps she could develop an easy mosquito-detector. So there she was in her lab at Stanford, with her mentor Manu Prakash, conducting the world’s unlikeliest recording session, with high-performance microphones that they had borrowed.

And then her cellphone rang.

The chiming made her wonder whether the phone could pick up a mosquito’s buzz, so she put in the cages, and found that it recorded the insects just as well as the fancy studio microphones. “We realized that people everywhere are walking around with mosquito detectors in their pockets,” she says.

A mosquito’s buzz reveals not only its presence, but also its identity. Each species seems to hum at its own distinctive pitch, and Mukundarajan and Prakash have shown that cellphone recordings are good enough to classify these insects. They’ve essentially created Shazam for mosquitoes. “I tell people there are 3,500 species of mosquitoes and they laugh,” says Prakash. “To them, it’s just a mosquito. We need to change that if we’re to fight these diseases properly.”

In poor countries where mosquito-borne diseases are especially rife, scientists usually monitor the insects by setting traps, in just a few houses within a given village. Teams of experienced technicians then laboriously sift through the insects, identifying them under a microscope. The resulting data is valuable and hard-won, but also patchy and limited. “There’s a huge and growing need to develop easy-to-use and low-cost strategies for doing mosquito surveillance on a mass scale,” says Heather Ferguson from the University of Glasgow, who thinks that Mukundarajan may have found one. “It an extremely innovative solution.”

Globally, there are 5.2 billion cellphone users, and the market is growing with particular vigor in Africa, Asia and Latin America—the parts of the world that suffer most from the bites of mosquitoes. And Mukundarajan says that even “really old flip-phone that one of us used in high school” could accurately record the insects in the lab. “That really convinced us that mobile phone-based surveillance could become a truly global tracking solution.”

Mukundarajan took her dated flip-phone to mosquito hatcheries at the Centers for Disease Control and Prevention (CDC) and recorded the noises of captive mosquitoes from 19 disease-carrying species. In most cases, she could identify them by sound alone. In a few cases where two species sound similar, time and place would provide the vital clue; for example, the malaria-carrying Anopheles gambiae overlaps in pitch with the Zika-carrying Aedes aegypti, but the former bites at night, while the latter is a day-biter. Only a few pairs of species were potentially indistinguishable; even very similar ones like Aedes aegypti and Aedes albopictus were easy enough to tell apart.

The phones don’t work at a distance. You need to get one within 5 centimeters of a mosquito to record it successfully. That might seem unfeasible, but you don’t need to chase a zigzagging insect through the air. “You spot a mosquito, wait for it to sit down, and bring the phone close to it,” says Prakash. “That single buzz when it flies off is all we need.”

The team did this successfully in the field. Even though a mosquito’s pitch changes with its age, size, and health, these variations are no bigger than the frequency range of a single individual’s flight. Put it this way: some mosquitoes may be natural sopranos and others may be natural tenors, but everyone has the range of the full choir. Which meant that Mukundarajan could successfully match wild, free-flying insects against the database she built at the CDC.

So could others. She and Prakash recruited an army of volunteers to test their technique. At the Big Basin Redwoods State Park in California, a team of 13 hikers used their personal phones to record around 125 mosquitoes in just 3 hours. The team repeated their efforts in a busy market street within Madagascar’s Ranomafana village. “These were naive volunteers,” says Prakash. “I gave them half-an-hour of training and left them on their own. Which is the point: Anyone should be able to do this.”

That’s a familiar refrain from Prakash. I’ve previously written about his Foldscope, a $1 pocket microscope that can be folded from a sheet of paper, and his Paperfuge, a 20-cent device modeled after a children’s toy, which can replace bulky, expensive lab centrifuges. For these inventions and others, Prakash earned one of the coveted $625,000 MacArthur genius grants in 2016. Turning phones into mosquito-detectors is just part of his quest to bring science to the masses.

The team have uploaded a paper describing their work to the bioRxiv server, and submitted it for peer review elsewhere. In the meantime, Ferguson notes that the technique isn’t quite ready for primetime yet. The team needs to show that they can tell the difference between many mosquitoes flying at the same time, against a range of background conditions. But she adds, “If an effective, phone-based tool for measuring mosquito activity like this were available, I’m sure myself and many  my colleagues would be very keen to use it.”

Other scientists are taking a similar approach. Marianne Sinka from the University of Oxford is designing an acoustic sensor that can automatically detect mosquitoes from afar. “Having citizen scientists chase round after mosquitoes introduces a massive bias in the sampling,” she says. They won’t capture the many species of mosquito, particularly malaria vectors, that have adapted to bite when their host is vulnerable—indoors and asleep.”

Sinka also notes that some phones aren’t suitable for recording mosquitoes, because their software removes a portion of the key tones. “But at the end of the day, accurate mosquito surveillance is essential,” she says. “If these guys can set up a successful citizen science project to get us closer to this goal, it can only be a good thing.”