Over the last few summers, John Keven has spent many long nights under the stars in Papua New Guinea. For 12 hours at a time, he’ll scour a giant green net set up between thatched huts, looking for resting mosquitoes every 20 minutes. When he spots one with his headlamp, he quietly approaches, extending a long rubber tube to suck the bug off the net. Then he blows it from the tube into a container for analysis—in a lab halfway around the world.

Keven is a doctoral student at Michigan State University, and leader of the mosquito-catching team. “Of course it is difficult to stay awake the entire night,” he says, “but I take coffee to help me.” It’s worth the effort: The undigested blood inside the Anopheles punctulatus mosquitoes Keven collects is going to the research team at the Cleveland Clinic in Ohio, which uses DNA markers to identify what the insects feed on through the night—information that could help predict how they spread disease.

The team’s recent testing, published last month in the journal PLoS Neglected Tropical Diseases, revealed that this type of mosquito feeds on a wider range of species than expected, potentially influencing the way it transmits malaria. The bugs feast on the humans in the villages, but also the pigs, dogs, mice and even marsupial species in the area. But this study is only one of a growing number of attempts to characterize mosquito behavior by analyzing the blood they suck.

With the recent emergence of Zika virus, entomologists say that matching the DNA fingerprint of human blood inside mosquitoes with individuals could help shed light on how these insects spread disease—and who is most vulnerable. “The extent to which mosquitoes don’t bite on everyone the same might actually be important when you think about who’s most important to vaccinate,” says Steve Stoddard, an entomologist at San Diego State University who has studied mosquito feeding behaviors. Data from this type of work could influence how researchers mathematically model the possible future spread of diseases carried by mosquitoes.

In 2014, Stoddard and his colleagues analyzed the feeding behaviors of the Aedes aegypti mosquito, the species that is a prime suspect in the current spread of Zika virus. This species can carry dengue, too, and it likes to hang around inside human dwellings, making it even riskier. The scientists collected mosquitoes from inside 19 households in Iquitos, a Peruvian port city on the Amazon, along with cheek swabs to capture DNA from 275 residents.

By comparing DNA signatures, Stoddard’s team matched the blood meals of 96 mosquitoes to specific individuals in the households. Turns out, people who stayed inside more often and the bigger members of the household were more likely to get bitten. “It’s relative to who else is available within the perceptual range of the mosquito, which kind of makes sense,” Stoddard says. “It’s the biggest steak on the counter, that’s the one the dog goes to.”

There are still too few studies matching mosquito blood meals to people to draw definitive conclusions about which humans they prefer. A 2003 study actually found hints that the A. aegypti mosquito might prefer young adults and males, while a 2014 study showed that people under 25 years of age were actually less likely to be bitten.

But while you’d probably love to know what makes you so tasty to a mosquito, global health researchers are more interested in understanding broad insect feeding patterns. DNA fingerprinting work confirms that the insects sometimes bite more than one person during a feeding cycle, and that could have implications for disease modeling. “If your model required 100 mosquitoes to be infected and every single mosquito fed on just one person, you have that one-to-one relationship and your model has boundaries,” says Peter Zimmerman, a Case Western biologist and global health researcher who helped lead the study that Keven worked on. “Now, if mosquitoes feed on four people at a time, suddenly the mosquito could be spreading its disease to four times the number of people.”

Analyzing the blood inside mosquitoes will continue to add detail to those models. Laura Harrington, who heads the entomology department at Cornell University, has studied the A. aegypti mosquito for almost two decades. Her DNA fingerprinting work in Thailand revealed that mosquitoes were feeding on people from outside the immediate community studied—most likely folks who were stopping in the villages during the day to sell produce and kitchen gear. There was even a truck that would stop by to sell a favorite snack of red ants. “Like an ice cream truck, it had a little loudspeaker on it and it would say ‘Red ants, red ants!’ as it drove through,” she says.

The need to understand those behaviors is all the more pressing with the current Zika outbreak. On Wednesday, the US Centers for Disease Control and Prevention officially confirmed that the virus can cause birth defects, including a condition known as microcephaly in which a baby is born with an unusually small head. Harrington hopes that DNA fingerprinting will prove useful for fighting Zika and other diseases like it. Perhaps, she says, by identifying whom mosquitoes avoid, in the future scientists might be able to develop odors that do a better job repelling these insects.

“Every bit of data we gather from mosquito behavior or virus host is of upmost importance in this war, which is far from over,” says virologist Davis Ferreira, deputy director of the Microbiology Institute at the Federal University of Rio de Janeiro in Brazil.

So Keven continues his bug-trapping work, despite the dangers. During his nightly excursions, he wears protective clothing that covers his whole body and uses bug spray, but he can’t avoid bites entirely. “I was bitten while I was out there collecting mosquitoes last summer and I got infected with two species of malaria,” he says. “That’s the risk involved in doing this type of work.”