Non-invasive saliva sampling: Grantee Maristela Martins de Camargo

July 11, 2025

As streets around the world emptied during the COVID-19 lockdowns, wild animals ventured into the silence.

From her home in São Paulo, Brazil, Maristela Martins de Camargo came across videos of animals exploring the deserted city streets. “Oh, my God!” she remembers thinking. “These animals are just around the corner, and we haven’t been looking at them.”

Maristela had always walked her dog in a particular park near the outskirts of São Paulo. But the sudden visibility of so many wild animals in cities around the world made her much more aware of the marmosets living in the park, and in the forests bordering the city.

As an Associate Professor of immunology at the Universidade de São Paulo, Maristela’s thoughts went straight to disease. What pathogens might the humans — and dogs — of the city be spreading to marmosets, and vice versa? How were the marmosets and other wild animals faring in this shared environment?

Connecting the dots

“I started reading the literature on pathogen circulation between people, domestic animals, and wild animals in the outskirts of cities,” Maristela says.

She soon saw that there was a great deal we don’t know — the data just wasn’t there. And the more she read, the more it became clear that sampling was the major bottleneck.

Biological samples were being collected from wild animals in invasive or distressing ways, and no one was collecting large datasets on pathogen presence because the methods were costly and required highly-trained animal handlers. “Epidemiologists weren’t modelling pathogen circulation in wild animals because they had almost nothing to work with,” Maristela says.

This was a pivotal moment for Maristela. She had always been concerned about animal welfare: Motivated by a desire to help animals, she originally trained in veterinary medicine. But the suffering she witnessed at the slaughterhouses she visited as a veterinary student pushed her into a career in immunology instead, where she made it her goal to elevate the scientific and ethical standards of the field by moving away from the animal models that she and her colleagues were trained to use.

Wild Animal Initiative grantee Maristela Martins de Camargo, DVM, PhD, is Associate Professor at the Universidade de São Paulo in Brazil, where she works in the Laboratory of Molecular Immunoregulation.

Photo courtesy of Maristela Martins de Camargo.

In her personal life, Maristela became an environmental activist, feeling that humans’ changed relationships with nature and nonhuman animals were affecting animal welfare, biodiversity, and the climate in alarming ways. But until now, she hadn’t found a way to use her scientific career to advance the same causes.

Then came the pandemic. And with it, the opportunity she had been searching for. “I found what I can do to help,” she realized. “Solve the sampling bottleneck.”

A novel method for non-invasive saliva sampling

Maristela had an idea for a saliva sample collector that would enable researchers to gather biological data on a large scale, without compromising the welfare of the animals being studied. 

She formed a collaboration with Dawn Zimmerman, a veterinarian, epidemiologist, and wildlife expert who was interested in how the saliva collector could support her work uncovering patterns of disease occurrence. Along with postdoctoral researcher Mariana Furtado, they applied for and received a grant from Wild Animal Initiative in 2024 to test the saliva collector and use it to gather welfare-relevant data.

Biological sampling is foundational to animal physiology research because key markers of animal health, genetics, and welfare can be measured in substances like saliva, blood, feces, and tissue. But collecting samples of those substances from wild animals is typically invasive (breaking the skin of the animal) or distressing (causing disturbance to the animal). Researchers often use traps, physical restraint, or sedation to physically handle animals, risking stress, injury, or even death. The process can also expose researchers themselves to injury and disease.

Some less distressing, non-invasive methods already exist, like swabbing bite marks on vegetation or offering chew toys with oral swabs. But these approaches are sometimes untargeted, not always effective, and might still be disruptive since they require researchers to be present at the study site.

This is where the device that Maristela and her colleagues are testing comes in: a 3D-printed box containing filter paper and polymer beads that have been rolled in human-food-grade artificial scent oils. It’s in a new category of baited, non-invasive sample collectors that minimize human presence in animals’ habitats.

Undergraduate student Aspen Gonçalves (left) and Mariana Furtado (right) install the collection device at a field site. Photo courtesy of Maristela Martins de Camargo.

Mariana Furtado sets up a wildlife camera to record animals interacting with the collector. Photo courtesy of Maristela Martins de Camargo.

The team uses a variety of scent oils to attract a variety of predefined species: some are partial to banana, while others prefer sugarcane. Rose, tuna, and Tutti Frutti are also on offer for more adventurous eaters.

The device is attached to a tree — sometimes up in the canopy, sometimes near the ground. Attracted by enticing smells, animals investigate the box and lick the filter paper, thinking it’s a tasty treat. Their saliva is collected on the paper, all while a nearby camera records the action. 

Each day, a researcher changes out the filter paper, bringing the sample and the footage back to the lab for identification. So far, the team has collected samples from two species of marmosets, two species of opossums, rodents, bats, and bees — who seem to be attracted only to particular scents, a preference that has never before been described.

Now that they’ve tested the collector, the team is at work on an automated model that will reset the filter papers for them each time saliva is collected. The goal is to reduce the frequency at which they have to pick up the papers by hand, so that they can keep their presence at the study sites to a bare minimum.

Collecting welfare-relevant data

Using the video footage from the wildlife camera, Maristela and her colleagues match the saliva to the licker and then analyze it for levels of hormones like cortisol and dopamine, which can offer insights about the animals’ welfare.

Their plan is to compare the hormone levels of animals living in two areas of the Universidade de São Paulo’s campus: a forest where there are only natural food sources and relatively little anthropogenic activity, versus an area near a restaurant, where animals are frequently hand-fed and forage in garbage cans. They hope that the comparison will give them an indication of which habitat better supports the welfare of the animals.

The information they glean could provide a foundation for future researchers to pinpoint which factors might be contributing to the local animals’ welfare: What role does human activity and its absence play? What about easy access to human food sources in urban environments, versus the experience of having to forage in the wild, where food is sometimes scarce?

But the project is also a major part of Maristela’s mission to solve the sampling bottleneck in epidemiology. If the method can help field researchers supply the large datasets that epidemiologists need, Maristela says, “they can make better predictions about where to intervene before disease outbreaks happen.”

If successful, the method could have broad applicability, enabling researchers in any field that studies wild animals to collect saliva samples in a non-invasive, non-distressing way. It would be a win for the animals being studied, and for the researchers, whose results could become more accurate and relevant by studying animals with as little human interference as possible.