The air above Dublin contains, among other things, genetic traces of cannabis, poppy, and magic mushrooms. Not the substances themselves, but fragments of their DNA, shed from plants, carried on particles, and pulled out of the atmosphere by a standard air filter running for a few hours. That finding comes from a study published on 3 June 2025 in Nature Ecology & Evolution by David Duffy and colleagues at the University of Florida’s Whitney Laboratory for Marine Bioscience, along with collaborators at the University of Copenhagen, Yale, the Barcelona Institute of Science and Technology, and the US Geological Survey.
The paper is titled “Shotgun sequencing of airborne eDNA achieves rapid assessment of whole biomes, population genetics and genomic variation” and it sits in a field that has moved considerably faster than most people outside it have noticed.
What environmental DNA is, and why air is a new frontier
Environmental DNA, abbreviated eDNA, refers to genetic material collected not directly from an organism but from what it leaves behind in its surroundings: shed skin cells, mucus, faeces, pollen, exhaled particles. For around two decades, eDNA monitoring has been a standard tool in aquatic ecology. Water from a river or lake can be filtered to capture DNA from every species that lives in or passes through it, and the technique has been used to detect rare fish, invasive species, and aquatic pathogens without ever catching or seeing the animals themselves.
Air as a medium for eDNA collection is newer. The first demonstrations that animal DNA could be reliably detected in outdoor air samples came in 2022, from two independent groups: one led by Elizabeth Clare, then at Queen Mary University of London, which sampled air at a zoo in England and detected DNA from the animals held there, including species the animals themselves had been fed; and a parallel effort led by Kristine Bohmann at the University of Copenhagen, which worked at a zoo in Denmark. Both groups used targeted approaches, looking for specific genes in specific species.
What the Duffy group has now demonstrated is a different approach at a different scale. Rather than targeting particular species with PCR probes, they used shotgun long-read sequencing: a method that reads whatever DNA fragments are present, without specifying in advance what you are looking for. Applied to air filter samples, this produces what is effectively a census of every organism whose genetic material happened to be present at the time of collection.
What the paper found, and how
The team collected airborne eDNA samples at two sites: a subtropical forest in Florida and locations in Dublin. The Florida samples were used primarily for wildlife and population genetics analysis. The Dublin samples, drawn from an urban setting, were used to assess pathogen surveillance and, as it turned out, to demonstrate one of the stranger capabilities of the method.
From outdoor air in Florida, the team performed population-level genetic analysis of two species: bobcats (Lynx rufus) and golden silk orb-weavers (Trichonephila clavipes). Using DNA captured from the air alone, without trapping, spotting, or sampling the animals directly, the researchers were able to place individual animals in their geographic population of origin. This is phylogenetic placement from ambient air, a capability that previously would have required physical specimens.
From Dublin, the team picked up DNA signatures from hundreds of human pathogens, including viruses and bacteria, floating in the city air. The drug-derived DNA — cannabis, poppy, fungi with psychoactive species — was a secondary finding, illustrating the breadth of the method: any organism with DNA, or any object carrying DNA, leaves a detectable trace in the air around it.
The paper also reports that a single researcher could run the full pipeline from air filter collection to completed species analysis in as little as two days, using compact portable sequencing equipment and cloud-based analysis software. That speed and accessibility is part of what the authors argue marks this approach as practically deployable rather than merely experimentally interesting.
The method compared to prior work
The distinction between the Duffy approach and earlier airborne eDNA work matters for understanding what is genuinely new here. Clare and Bohmann’s 2021–2022 zoo studies used short-read targeted sequencing, looking for specific gene regions from known species lists. Reliable, and an important proof of concept, but limited to what you already know to look for. Shotgun long-read sequencing reads the whole genome rather than just targeted fragments, and does not require a predetermined target list. This means unknown species, unexpected pathogens, and organisms that standard probes would miss all show up in the data.
A 2025 study from Sweden, not by the Duffy group, demonstrated that airborne eDNA captured and stored on archive filters from a radionuclide monitoring station could yield retrospective ecological data stretching back three decades, with DNA preserved in filter material that had never been intended for biological use. That result points to a further implication of the field: the world has been inadvertently collecting airborne eDNA in existing environmental monitoring infrastructure for years.
The Duffy paper adds to this an explicit demonstration of population genetics at the level of individual animals, viral variant calling from urban air, antimicrobial resistance gene surveillance, and whole-biome species assessment, all from a single filter sample. No single prior paper had combined those capabilities across one study.
The drug detection result in perspective
It is worth being clear about what the Dublin drug DNA result does and does not show. The filters captured DNA from plants associated with cannabis, poppy, and psilocybin mushrooms. They did not detect the psychoactive compounds themselves. DNA from cannabis plants does not indicate that cannabis smoke was present, or that anyone nearby was using the drug. It indicates that organic material from those plants, carrying their genetic signature, was present in the air at some concentration sufficient to be captured and sequenced.
Dublin has legal uses of poppy derivatives in pharmaceutical supply chains, legal hemp cultivation, and, like many European cities, ambient urban biodiversity that includes plant material from many sources. The finding is interesting as a demonstration of the method’s sensitivity, and it raises real questions about surveillance applications that the authors themselves flag. It should not be read as evidence of widespread drug use in the city or as a monitoring tool ready for law enforcement deployment. The paper is classified as an observational study; its method is research-grade and requires significant analytical infrastructure.
The surveillance problem the authors raise themselves
Duffy’s group has a track record of pressing the ethical implications of eDNA before others in the field have caught up. An earlier paper from the same lab, in 2023, demonstrated that human DNA — specific, genetically identifying human DNA — is present in ambient environmental samples from water, soil, and sand in publicly accessible places. At the time, they called for ethical frameworks for the collection and use of such data, noting that the same techniques used to identify a bobcat’s population of origin can identify a person’s genetic ancestry, health predispositions, and family relationships from material they shed without consent or knowledge in a public space.
That concern is amplified in the current paper. Shotgun sequencing of urban air picks up human genetic material routinely. The paper demonstrates this with samples from Dublin and Florida and notes that exome capture, a technique that targets the protein-coding portions of the genome, was successfully applied to air filter samples. The paper states that this capability reveals “the future feasibility” of applications the authors describe as requiring governance frameworks that do not yet exist.
This is not an abstract concern. Airborne eDNA sampling does not require proximity or contact. A monitor running for hours or days in any indoor or outdoor space will accumulate genetic signatures from everyone who passes through it. The paper is careful to raise this and to position the research community as having a responsibility to engage with the question before applications outpace policy. That framing is rare enough in methods-demonstration papers that it warrants noting.
What to watch next
The immediate research directions the paper opens are in pathogen surveillance and biodiversity monitoring. Both have clear institutional demand: public health agencies want early-warning systems for emerging respiratory pathogens, and conservation bodies want faster, cheaper ways to census wildlife populations in large remote areas. The two-day pipeline the Duffy group demonstrates, running on portable equipment and cloud analysis, is within reach of field deployment in both contexts.
The governance question will take longer. Existing legal frameworks for genetic privacy were designed around biomedical contexts: clinical samples, databases, insurance. Ambient air sampling fits none of those categories cleanly. Duffy’s lab has been arguing since at least 2023 that this gap needs to be closed before the technology is deployed at scale by actors with less interest in flagging the problem.
Whether the field listens is a different question. Air quality monitoring infrastructure already exists in thousands of cities. The Swedish archive study shows that usable eDNA has been sitting in those filters for decades. The question is not whether airborne environmental DNA is real, or whether it carries enough information to be consequential. The paper establishes that clearly. The question is who decides what it is used for.
