What do a cat, a dingo, and a goanna have in common? It’s in the iDNA

A cat, dingo and goanna in Namadgi National Park were the latest animals recorded via DNA from an insect’s belly. Tim Cutajar at the Australian Museum and Dr Stephanie Pulsford tell us how!

If you’ve ever been out walking in nature, you know some species are not all that easy to spot – what you see is often just a small fraction of what’s out there. How many times have you seen an exquisite bird or rare marsupial, but only on the information signage in a national park? This can happen even to scientists studying and managing those species. We know why this is; some species are secretive or only active at certain times, others may be excellent at camouflage, and some are just plain rare. Knowing why we can’t find a species is a good start, but we also need techniques to make finding them easier.

Recently, we developed a technique for detecting rare frogs using their parasites. There are tiny, blood-sucking flies that feed only on frogs – they find these frogs by following the sounds of their calls. We set mosquito traps playing frog call audio along rainforest streams at multiple forest sites in NSW and caught hundreds of these frog-biting flies. We sequenced the DNA from the flies’ blood meals and detected frog species that we’d not seen or heard out in the forest. This was great news for rainforest frog research.

This new frog detection technique fits into what’s broadly called invertebrate-derived DNA (iDNA). Other iDNA studies have used leeches and blowflies and detected other rare animals. Like our study, most were performed in rainforest and/or tropical areas, but we wanted to know if our frog iDNA technique would work in other environments.

At the end of 2020, we ran a trial of the technique in the seasonally cold, dry forest of Namadgi National Park in the Australian Capital Territory. We ran frog call traps overnight at a few sites scattered around the park and returned in the mornings to see what had been collected. We’d caught many flies, but none that we recognised as frog-biting species. Disappointed, but hopeful that perhaps we’d misidentified some true frog-biting flies in our catch, we decided to go ahead with DNA extractions from all of the flies and check them for vertebrate animal DNA.

We were astonished by what we found. Three flies yielded high-quality DNA that matched to vertebrates. Excited to see what frog species we’d detected, we analysed the first sequence … cat – very strange. The next one? Dingo. By the third sequence, we really had no idea what to expect. We would never have guessed it, but it was from a species of goanna, the Rosenberg’s monitor (Varanus rosenbergi).

There are a few possible reasons why we didn’t detect frogs; maybe there are no frog-biting flies in the cool, dry forests of southern Australia, or if there are, maybe they prefer different habitats to where we surveyed. It’s also possible catastrophic wildfires that tore through the area the previous year simply wiped them out – we just don’t know. But our disappointment was eclipsed by our confusion at the eclectic mix of species we did detect.

Our confusion was rooted in a simple fact – flies that parasitise medium to large-bodied animals that don’t vocalise reliably like frogs (i.e. cats, dogs, and lizards) tend to use chemical cues to find their hosts. So how did we collect such flies when the only ‘bait’ we used was frog call audio? The most logical answer is this: coincidence. We think these flies just happened to fly close enough to our traps to get sucked in, and the implication here is huge. Either we were extremely lucky or, more likely, these flies are very abundant in the area. What that would mean is that, with an attractant more appropriate to non-frog specialist flies, this technique could prove extremely useful for the detection of key species in seasonally cool, dry Australian forests.

Rosenberg’s monitor is threatened in some areas and is the subject of ongoing monitoring in Namadgi. Dingoes are common in the area, and are widely studied by evolutionary biologists, making DNA samples important and in high demand. Other actively managed and elusive species in the area in need of improved detectability include the broad-toothed rat (Mastacomys fuscus), alpine skink (Cyclodomorphus praealtus), and the highly secretive Canberra grassland earless dragon (Tympanocryptis lineata).

Our detection of a cat through iDNA is also significant. It highlights iDNA’s potential for early detection of introduced species. Feral cats are widespread in Australia. Although detecting one in Namadgi confirms what we know, there are fenced wilderness areas that intentionally exclude cats and islands where they have never established. In places like these, routine iDNA surveys could be key to early detection of breaches and allow timely intervention.

It turns out iDNA could be an effective tool in cool, dry southern Australian forests where it has never been used before. We could detect threatened and rare species better by adding iDNA techniques to their monitoring strategies, and do the same for invasive predators in predator-free, naïve animal populations for early intervention. The next step is to test these ideas. It’s our hope that some minor strategic tweaks can lead to major improvements in our knowledge of southern Australia’s imperilled biodiversity, and our ability to conserve it.