Frozen Squirrel Droppings Yield 700,000 Years of Ancient DNA — Including Woolly Mammoths

The oldest environmental DNA ever recovered from animal feces sat for hundreds of thousands of years in permafrost beneath Canada's Klondike region, sealed inside the burrows of Arctic ground squirrels no bigger than a loaf of bread. When researchers finally cracked open the frozen pellets — each about the size of a rabbit dropping — what they found was, in the words of lead author Tyler Murchie, "spectacular" [1].

The study, published June 9, 2026 in Nature Communications, reports the reconstruction of 18 mitochondrial genomes from 13 coprolites (fossilized feces) collected from ground squirrel burrows in central Yukon [2]. The samples span from roughly 17,000 to approximately 700,000 years ago, making them among the oldest sequenced sources of ancient environmental DNA on record [3]. Among the genomes: six woolly mammoths from different time periods, along with steppe bison, horses, and a dozen ground squirrels [4].

"I can't believe that we were able to get these insights from squirrel faeces," Murchie, a paleogenomics researcher now at the Hakai Institute in British Columbia, told reporters [1].

How Squirrel Poop Became a Genetic Archive

The Arctic ground squirrel (Urocitellus parryii) is not a picky eater. During its roughly four-month active season, it consumes plants, fungi, insects, and — critically — carrion, including the remains of large mammals [4]. Researchers have documented squirrels eating everything from whale blubber to rodent carcasses, leading one Nature headline to call them "zombies of the Pleistocene" [5].

This scavenging behavior is what makes their feces so scientifically valuable. When a squirrel gnawed on the carcass of a dead mammoth, fragments of mammoth DNA entered the squirrel's gut and were deposited in its feces. The squirrel then retreated into its burrow — sometimes at depths where permafrost permanently sealed the chamber — and the droppings froze in place, preserving genetic material from across the ecosystem [3].

"They're just trying to eat everything they can," Murchie explained. "It kind of makes a perfect ancient DNA repository" [6].

The burrows also contained other biological material — nuts, seeds, leaves, bones, and fur — packed in by the squirrels during their active months [1]. When researchers inserted fluid to extract genetic material, the samples produced an "overwhelming" smell, confirming they had never mineralized like dinosaur coprolites but remained organic matter, frozen in time [7].

What the DNA Revealed

The team identified DNA from a wide range of Pleistocene fauna and flora. Beyond the six mammoth mitochondrial genomes, the coprolites contained genetic traces of grey wolves, a large cat (either a cougar or American cheetah), lemming, pika, caribou, snowshoe hare, and extinct species of bison and horse [2][4]. More than 200 plant groups were also identified, along with invertebrates including parasitic worms [2].

The 18 reconstructed mitochondrial genomes broke down as follows: 12 ground squirrels, three horses, two bison, and one hare — assembled alongside the mammoth and other environmental DNA sequences recovered at lower coverage [3].

One of the study's most striking findings had nothing to do with mammoths. The researchers discovered a previously unknown genetic lineage of Arctic ground squirrel dating back 700,000 years — a lineage that no longer exists in the Yukon but whose closest living relatives are found only in western Siberia [2]. Until this study, fossil ground squirrel remains from that period in central Yukon had been assumed to belong to the same species found there today [2].

"We're able to pull things out that I never thought we could do," said Hendrik Poinar, Director of the McMaster Ancient DNA Centre and one of the study's senior authors [6].

Placing the Mammoth DNA in Time

The radiocarbon dates and stratigraphic positions of the coprolites — calibrated using layers of volcanic ash with known ages — place the mammoth DNA across a broad temporal window [7]. With samples ranging from 17,000 to 700,000 years ago, the mammoth genomes span multiple glacial and interglacial periods in Beringia, the ancient landmass connecting Asia and North America [4].

The last confirmed population of woolly mammoths survived on Wrangel Island, a remote spit of land off the Siberian coast, until approximately 4,000 years ago [8]. The Yukon mammoth DNA from this study predates that final population by thousands to hundreds of thousands of years, placing it within the period when mammoths were widespread across northern latitudes. Separate research by Murchie and colleagues, published in Nature Communications in 2021, had already used sedimentary ancient DNA from the same Klondike region to suggest that mammoths survived in the Yukon until roughly 9,700 years ago — some 3,000 years later than the fossil record had indicated [9].

The new coprolite-derived mammoth genomes could help fill temporal gaps in the mammoth genomic record. Scientists have previously sequenced DNA from mammoth teeth over a million years old [10], and a 2024 study reconstructed a complete 3D woolly mammoth genome from freeze-dried skin [11]. But mitochondrial genomes from datable fecal sources offer a different kind of evidence — one tied to a specific ecological context rather than a single specimen.

Source: Nature, Science, Cell (compiled)

Data as of Jun 10, 2026CSV

DNA Preservation: How Coprolites Compare

Ancient DNA research has advanced rapidly over the past decade, with permafrost emerging as the gold standard for long-term genetic preservation. The field has produced a series of landmark recoveries: DNA from a 700,000-year-old horse leg bone in 2013, mammoth teeth exceeding 1.2 million years in 2021, and two-million-year-old genetic material from Greenland sediments in 2022 [10][12].

Source: OpenAlex

Data as of Jan 1, 2026CSV

Coprolites occupy an unusual niche in this landscape. Bone and tooth samples typically yield higher concentrations of endogenous DNA — genetic material from the organism itself — because the hydroxyapatite mineral matrix protects DNA strands from enzymatic degradation [13]. Fecal samples, by contrast, are rich in microbial DNA from the gut microbiome and contain a mixture of host DNA, dietary DNA, and environmental contaminants [13].

This makes coprolites less reliable for recovering deep, high-coverage genomes of a single species, but more valuable as biodiversity proxies — snapshots of entire ecosystems captured in a single deposit. The ratio of squirrel gut microbiome DNA to ingested environmental material in these samples has not been publicly quantified, though the breadth of species identified — mammals, birds, invertebrates, plants, microbes — suggests the environmental fraction is substantial [3].

Murchie acknowledged limitations: "Some of the DNA may have been picked up from the coprolite's surface at a later time," and incomplete genetic databases may affect species identification [3]. The study's methodology involved developing custom molecular baits and reducing sample quantities to overcome inhibition from organic compounds in the feces — a technical hurdle that had previously limited coprolite DNA work [6].

The Contamination Question

Skeptics of environmental DNA research have long argued that ancient DNA in sediment or feces can be contaminated by post-depositional leaching — genetic material seeping into a sample from surrounding soil over millennia, producing false positives about which species were actually present at a given time [14].

This concern is especially relevant for permafrost samples, where DNA from organisms buried in adjacent soil layers could theoretically migrate through freeze-thaw cycles or groundwater movement. Standard protocols for addressing this include applying chemical or DNA-based tracers to sediment cores to detect infiltration, running parallel extractions with sediment-free blank controls, and using bioinformatic filtering to distinguish endogenous from contaminant sequences [14].

The published reports on this study do not detail the specific contamination controls used, beyond Murchie's acknowledgment that surface contamination remains a possibility [3]. The sealed nature of the permafrost burrows — some of which had been continuously frozen for hundreds of thousands of years — provides a degree of natural protection against leaching, but the absence of publicly described authentication protocols is a gap that peer reviewers and outside commentators may scrutinize.

Mikkel Pedersen of the University of Copenhagen, a collaborator on the study, has published extensively on sedimentary ancient DNA authentication methods [15]. His involvement suggests the team applied rigorous controls, but the specifics will matter to the broader ancient DNA community, which has been burned before by contamination-driven false claims.

De-Extinction and the Colossal Question

The discovery arrives at a moment when woolly mammoth genetics is no longer a purely academic pursuit. Colossal Biosciences, a Dallas-based biotechnology company, has raised $555 million to pursue the "de-extinction" of the woolly mammoth by engineering mammoth-like traits into Asian elephant cells [16]. The company has used 65 different mammoth genomes in its computational work, analyzing 59 woolly, Columbian, and steppe mammoth genomes ranging from 3,500 to over 1.2 million years old [17]. In March 2025, Colossal announced it had created gene-edited "woolly mice" with mammoth-inspired hair traits, a proof-of-concept step on the path to producing mammoth-elephant hybrid calves by 2028 [16].

Does the squirrel coprolite find change Colossal's timeline or feasibility? Probably not in a material way. Murchie himself noted that Colossal "already have so much DNA to go off of" [1]. The mitochondrial genomes recovered from the coprolites — while scientifically valuable for phylogenetic analysis and population genetics — represent a small fraction of the total genomic information needed for de-extinction engineering, which requires high-coverage nuclear genome data [17].

Where the coprolite DNA may prove more significant is in resolving taxonomic questions. Genomic research has confirmed that at least two distinct lineages of woolly mammoth coexisted in Siberia, a finding that paleontologists had initially resisted because skeletal morphology alone did not show sufficient variation [10]. Mitochondrial genomes from six mammoths across different time periods, recovered from a single geographic context, could add data points to this debate — though whether the coprolite-derived sequences cover the specific genomic regions needed to address the steppe-versus-woolly mammoth question remains to be demonstrated in follow-up analysis.

Murchie stated that the team's genetic data would be publicly available [1]. The question of whether recovered mammoth genomic data can be used for proprietary de-extinction projects — versus remaining in open scientific databases — is governed more by institutional data-sharing policies and publication norms than by any specific legal framework. Colossal stores approximately 3.8 petabytes of genomic data through its computational biology spinoff, Form Bio [17]. Some Colossal-affiliated research has been published under Creative Commons open-access licenses, but the company's commercial interests create an inherent tension with the open-science ethos that has historically governed ancient DNA research.

Indigenous Rights and the Permafrost Economy

The coprolites in this study were collected from Yukon's Klondike region — land that falls within the traditional territories of several Yukon First Nations. The published coverage of this research does not mention whether indigenous communities were consulted before excavation or what legal frameworks governed the collection and export of samples for analysis [1][2][3].

This absence is conspicuous given the broader context. In Siberia, where mammoth remains are found in similar permafrost environments, the ivory trade has become a significant economic driver for indigenous Yakut communities. The Republic of Sakha (Yakutia) banned the export of mammoth tusks longer than three meters in an effort to bring transparency to the trade and reduce permafrost damage, though critics argue the restriction has fueled smuggling and harmed indigenous livelihoods [18].

In Canada, the legal landscape is different but the ethical questions are analogous. Yukon First Nations have self-governance agreements that grant varying degrees of authority over heritage resources found on settlement lands [19]. Whether ancient fecal pellets from ground squirrel burrows fall under the same heritage frameworks as, say, human archaeological remains is an open question — but the principle that indigenous communities should have a say in research conducted on their traditional territories is increasingly recognized across the sciences.

The researchers involved — from McMaster University, the Hakai Institute, and the University of Alberta — work within Canadian institutional frameworks that typically require ethics review for research involving indigenous lands or peoples. Duane Froese of the University of Alberta's Permafrost Archives Science Laboratory, who contributed samples from approximately 200 ground squirrel coprolites held in frozen storage, has worked extensively in the Yukon [6]. But the specific terms of community engagement for this study have not been publicly documented.

What This Find Actually Means

The significance of this research is not that scientists found mammoth DNA — mammoth genomes have been sequenced repeatedly from teeth, bones, skin, and sediment over the past two decades. The significance is the source and the method.

Ground squirrel coprolites represent a new category of ancient DNA archive: biological samples that capture not just one organism but an entire ecosystem's genetic signature, preserved by the accidental behavior of a small rodent and the geological accident of continuous permafrost. The roughly 200 coprolite samples still in frozen storage at the University of Alberta represent an untapped reservoir of Pleistocene biodiversity data [6].

The field of ancient permafrost DNA research has grown substantially, with over 3,500 papers published since 2011, peaking at 463 publications in 2023 [20]. This study adds a methodological tool — coprolite-derived environmental DNA — that complements existing approaches using bones, teeth, sediment cores, and ice cores.

Whether this tool will prove as reliable as its more established counterparts depends on answers to questions this study raises but does not fully resolve: How much of the DNA is genuinely from the squirrel's diet versus post-depositional contamination? Can the mitochondrial sequences be extended to nuclear genomic data? And will the communities on whose land these samples were collected have a voice in how the data is used?

The $2.3 million recently awarded to McMaster University and the Hakai Institute for related climate research using ancient DNA analysis suggests the work will continue [2]. The frozen pellets, it turns out, have more stories to tell.