Mosquitoes can be deadly, transmitting malaria, dengue, and Zika. But tracking them is tough. Now, researchers—led by bioengineer Manu Prakash of Stanford University in Palo Alto, California—have developed a new, cheap way to monitor these insects with mobile phones and a Shazam-like app that tells them apart based on their “songs.” Most mosquito species can be identified by the frequency of their wingbeats, so the web app—aptly named Abuzz—lets users upload recordings of mosquito sounds, identify the species, and map its location. But getting a good recording can be tricky: The insect cannot be more than 10 centimeters away from the microphone, and background noise cannot be louder than light traffic.
Prakash chatted with Science about this project shortly after the app launched on 31 October. This interview has been edited for clarity and length.
Q: Why did you start this project?
A: When I started my own [lab] group 6 years ago [at Stanford], the construction of the lab was going to take a long time, so I decided to travel and ended up in Thailand where I accidentally met a group of medical entomologists. One day, I walked into a room where a lot of researchers—who had collected thousands of mosquitoes from the field—were hunched over microscopes, counting [the insects] one by one. I wanted to build a more efficient and less labor-intensive system that could allow us to monitor mosquitoes at a much larger scale.
Q: What was your greatest “aha” moment?
A: Up to now all the acoustic work that's been done with mosquitoes has been with expensive microphones. But we realized that even old $20 flip phones are able to record mosquito sounds—and they have similar performance to fancy phones. Also, phones not only record the acoustic signature [of the mosquito], they also tell when and where it was recorded, which helps identify and locate the mosquito species.
Q: How does Abuzz work?
A: Most phones already have a sound recorder, so you just pull it out, identify where the phone’s microphone is, and then bring it close to the mosquito. As it flies, you get a buzz that you upload on the website. Then an algorithm matches it with a data set of 20 mosquito species that we’ve collected, and tell[s you] what species that mosquito most likely is. Right now, people have to upload the data on the web, but we're working on an app that makes this much easier to do.
Q: What were the biggest challenges?
A: Mosquitoes are phenomenal organisms, but they're quite challenging to work with. When I first arrived in Palo Alto, I was looking for a site where I could find lots of mosquitoes, and I found a swamp. The next day I told my students to come with me to record mosquitoes and we were there for hours for 2 days. Three or 4 days later, we started noticing blisters all over our bodies. It turned out we had walked into a forest of poison oak. … Now, we are all very good at spotting poison oak!
Q: You also tested your monitoring system with high school teachers in Madagascar. What was their reaction?
A: They couldn't believe it. When you hear the annoying buzz of a mosquito, [it] is because it's very close to your ear, but with a cellphone you can record things at distances of 5 to 10 centimeters. The second reaction [was] excitement, when people [realized] that when they play the sound of [two species], they can hear that the frequency for one of them is lower than the other. To drive this point home, we took the 20 species in our database and we collaborated with a musician to make a song where every key is the sound of a different mosquito species (you can listen, above).
Q: How do you think building global maps of mosquito distribution will help control efforts?
A: Mosquito surveillance is very expensive. If we can bring the costs down, countries that have no surveillance infrastructure might for the first time have the capacity to do surveillance. A couple of weeks back, I was in [New] Delhi and there was a peak of dengue that appears every year. So I started looking for how people monitor it, and I found out that the Delhi government pays community workers to go around and collect mosquitoes. Now imagine if a much larger community could engage with this: Instead of spraying the entire city of Delhi [with insecticides], one could use the mobile surveillance data to identify where the [disease-carrying] mosquitoes are. And it might be possible to curb the peak of dengue before it appears.
Q: How many people have contributed to this project so far?
A: Since we’ve launched Abuzz, we've gotten around 200 recordings. But we are starting to make posters that we'll share with schools, colleges, and universities around the world to engage people. It’s important that communities play a role because it's their kids’ lives that are being affected by these diseases.