What the Study Found
- Nuclear genomes place the American cheetah as a sister species of the puma, splitting from that lineage about 2.6 million years ago.
- Yukon fossils extend the species’ known range more than 20 degrees of latitude north, from Florida to Arctic steppe-tundra.
- Nitrogen isotopes put the Yukon cats at trophic position 4, about 6 permil above local carnivores and consistent with fish.
- Effective population size fell from nearly 100,000 animals to under 10,000 by 70,000 years ago, with no sign of inbreeding.
A LEFT metacarpal, a type of hand bone, came out of the frozen muck at a placer mine in the Klondike goldfields back in 2012. Nuclear genomes recovered from four fossil specimens show that the extinct predator long called the American cheetah was no cheetah at all, but a sister species of the puma. Three of those specimens came from the Arctic Yukon, hundreds of kilometres north of anywhere the animal was supposed to have lived, and the nitrogen locked into their collagen sits higher than that of any other carnivore of the same age in the region. The most economical reading of that chemistry is fish.
For decades this animal ran a very specific errand in North American paleoecology. It was the reason the pronghorn is fast, the ghost predator whose pursuit supposedly carved out the continent’s quickest hooves.
The Ghost That Made the Pronghorn Fast
That story rests on a body plan: slender build, elongated forelimbs, roomy nasal passages, the whole cursorial kit a sprinting cat needs. Mitochondrial DNA had hinted since 2005 that the resemblance was skin deep, but mitochondrial trees mislead in cats, which hybridise and swap lineages readily enough to scramble the signal. So Molly Cassatt-Johnstone, a doctoral candidate in the Paleogenomics Lab at the UC Santa Cruz, and her colleagues went after the nuclear genome instead, generating 22 fold coverage for a radiocarbon dated Wyoming individual and 38 fold for the best preserved Yukon one. The tree came back unambiguous: Miracinonyx trumani sits beside Puma concolor, and the two lineages parted roughly 2.6 million years ago.
Which leaves the cheetah likeness as convergence, two unrelated cats arriving at the same silhouette because open country rewards it. Their last shared ancestor with the African cheetah lived about 4.7 million years ago.
The Yukon fossils do something else to the story. They stretch the species’ known range by more than 20 degrees of latitude, from present day Florida to Arctic steppe-tundra, and drop it into a Late Pleistocene carnivore guild already crowded with cave lions, scimitar cats, gray wolves, brown bears and short-faced bears. Somewhere in that crowd, a medium sized cat found room.
“These cats were remarkably flexible, much like pumas are across their range today,” says Cassatt-Johnstone. Matthew Wooller, at the University of Alaska Fairbanks, puts the two ends of the range side by side: “They’re demonstrating uber-specialization at two ends of their range, while also feeding on two completely different food sources.”
What the Nitrogen Was Trying to Say
Nitrogen isotopes can reveal where an animal fed in the food chain during the last few years of its life. The heavier isotope, nitrogen-15, becomes more concentrated at each step up the food chain. In the northern cats, bulk nitrogen isotope values were about 6‰ higher than those of other carnivores living in the same region, and higher still than those of their likely land-based prey. That creates a puzzle: no known terrestrial prey on the Yukon steppe had nitrogen values high enough to explain the cats’ unusually strong signal.
Aquatic food chains offer one possible explanation. They can extend through more trophic levels, and anadromous fish such as salmon can carry elevated nitrogen isotope values inland from the sea.
The researchers tested that idea with compound-specific isotope analysis of individual amino acids. By comparing nitrogen isotopes in phenylalanine and glutamic acid, they could estimate the cats’ trophic position more directly. The Yukon animals came out at about trophic level 4, consistent with tertiary consumers. Their Wyoming counterparts averaged about 3, plus or minus 0.3, roughly what would be expected for a predator feeding mainly on herbivores.
The authors are careful here, and it is worth following them. Bulk isotope values on their own cannot resolve trophic relationships, they write, and whether these cats caught fish in the shallows the way fishing cats do, waited out the spawning runs, or simply scavenged spent carcasses off the gravel bars is not a question collagen can settle.
The genomics carries its own asterisk. Of 48 genes flagged as evolving under positive selection, none survived correction for multiple testing, and the loss of function results, including a broken sour taste receptor shared across every felid in the dataset, rest on two individuals.
A Long, Shallow Decline
The demographic history is the other finding, and possibly the sadder one. Effective population size for both populations slid from nearly 100,000 animals to under 10,000 by 70,000 years ago, and then stayed down there. What the genomes do not show is a bottleneck, or long runs of homozygosity, or any of the scarring that usually precedes a collapse. Just a long, shallow slide.
Beth Shapiro, who supervised the project, reads the slowness itself as the trap; a species that had been small for a million years had nothing left in reserve when the climate turned. Which raises an awkward question about the other bones in the other drawers, sorted by eye, filed under the wrong genus, waiting for someone to sequence them.
Reference
Cassatt-Johnstone, M., Wooldridge, T. B., Ghosh, S., Hall, E., Hewitson, S., Meachen, J., Wooller, M. J., Zazula, G., & Shapiro, B. (2026). Range and diet diversity in the Pleistocene American “cheetah,” Miracinonyx trumani. Current Biology. https://doi.org/10.1016/j.cub.2026.07.072
- Study type: Peer-reviewed report in Current Biology. Comparative paleogenomic and stable isotope study of radiocarbon dated fossil specimens, with phylogenetic, demographic and selection analyses.
- Sample size: Six Miracinonyx trumani fossil specimens. Four were sequenced and radiocarbon dated (three from Yukon, one from Wyoming) and two further Wyoming specimens were used for isotopes only, compared against genomes from nine other felid species.
- Material examined: Fossil bone from Quartz Creek and Bluefish Caves in Yukon Territory, Canada, and from Natural Trap Cave, Wyoming. Collagen was extracted for dating and isotope work, and bone powder for ancient DNA.
- Dating method: Radiocarbon dating of extracted collagen by accelerator mass spectrometry, calibrated against the IntCal20 northern hemisphere curve.
- Period covered: Specimens span roughly 17,000 to 34,000 years ago. The three Yukon fossils date to 31–34 thousand years ago and the sequenced Wyoming individual to 23.5 thousand years ago.
- Funding / conflicts of interest: National Science Foundation, Howard Hughes Medical Institute, and an Institutional Development Award from the National Institute of General Medical Sciences. Declared: senior author Beth Shapiro is employed by and holds stock options in Colossal Biosciences.
- Data availability: Raw sequencing data are in NCBI BioProject PRJNA1413399, four mitochondrial genomes in GenBank, and extended tables in the Dryad Digital Repository. No new code was generated.
- Main limitation: Bulk isotope values alone cannot resolve trophic relationships, so anadromous fish is an inference rather than a demonstration. Nuclear conclusions rest on two high coverage individuals, and no gene in the selection scan survived false discovery rate correction.
FAQ
Could a cat like this actually catch a fish?
A cat like this could plausibly have got hold of fish in several ways, and the study does not settle which. Living fishing cats hunt in shallow water and jaguars take fish directly, so the behaviour is well within reach of a medium sized felid. The Yukon animals may equally have scavenged spawned out carcasses along riverbanks or gathered at spawning runs when fish were dense and easy.
Does this change why pronghorn are so fast?
It complicates the usual explanation for why pronghorn are so fast. The idea that their speed evolved to outrun a cheetah-like pursuit hunter rested on the assumption that Miracinonyx hunted like a cheetah, and the genomes place it with pumas instead. Earlier morphological work had already pointed the same way, suggesting a generalist predator rather than a specialist sprinter.
How can a fossil bone reveal what an animal ate?
A fossil bone reveals diet through the chemistry of its collagen. Nitrogen comes in a heavier and a lighter form, and the heavier one concentrates a little more with each step up a food chain, so an animal eating predators carries a different signature from one eating grazers. Analysing individual amino acids rather than the bone as a whole sharpens that reading into an estimated position in the food web.
Is the fish diet actually proven?
The fish diet is not proven, and the authors say so directly. What the isotopes establish firmly is that the northern cats fed a full step higher up the food chain than the southern ones, higher than any other carnivore of the same age in the region. Aquatic prey is the most economical explanation for values that high, but bulk isotope readings on their own cannot identify a prey species.
Why does a broken taste gene matter here?
A broken taste gene matters here mostly as a curiosity about cats in general rather than about this species. The inactivated sour taste receptor turned up in every felid the team sampled, not just in Miracinonyx, so it is not an Arctic adaptation. Cats were already known to have lost sweet taste, and losing the gene does not necessarily remove the ability to detect sour flavours.
Cite This Page