A deep dive on my issues with Colossal’s so-called dire wolves
I had a pretty relaxing morning on April 7, 2025. I got up, fed the cats, made myself some tea, and attended a mindful writing session. I took a break for lunch before I was scheduled to attend a talk on informing environmental decision-making, just enough time to watch the previous night’s episode of Last Week Tonight with John Oliver. Then, my postdoc advisor Dr. Danielle Fraser texted me saying, “Seeing all this dire wolf stuff?”
I went on Bluesky and typed “dire wolf” into the search bar, hoping to find a cool new study. Instead, I found a bunch of scientists and science journalists up in arms about a number of news articles. “The dire wolf is back” claimed The New Yorker. Time Magazine proclaimed “The Return of the Dire Wolf.” Carl Zimmer in The New York Times was a little more reserved with an article titled “Scientists Revive the Dire Wolf, or Something Close.” Colossal Biosciences, a billionaire-owned “de-extinction” company, claimed it had revived the dire wolf, an extinct member of the dog family Canidae that went extinct at the end of the Pleistocene 11,700 years ago.
Here’s the thing: Colossal absolutely hasn’t “de-extincted” the dire wolf. And what they have done - and plan to do - is worrying. Let’s talk about it.
First, a disclaimer. I am a palaeontologist and evolutionary ecologist, but I am not a geneticist. I have dome some work with DNA for phylogenetics and have attempted to extract ancient DNA from a fossil (it didn’t work). I also think a lot about ethics, but I am not a bioethicist. This post is based on my understanding of these topics, but I am more than happy for experts to correct me on anything I have misunderstood. I will update this post accordingly.
Next, I think it’s important to note that none of this work is published or peer-reviewed. Without any technical documentation on the processes Colossal has used, we can only go off of what they have said publicly. Without any of this information being published and peer-reviewed, we have no way of knowing if they’re accurately interpreting what they’ve found in the dire wolf genome.
Since we don’t have a peer-reviewed, or even preprint, paper to go through, I’m going to walk through the claims on Colossal’s webpage for the dire wolf project, and supplement with interviews and news articles where needed.
Why Colossal thinks we need de-extincted dire wolves
One of the first things that is said on the webpage is:
Our team is proud to return the dire wolf to its rightful place in the ecosystem.
This starts us off with a doozy: what is the dire wolf’s rightful place in today’s ecosystem? I would argue that it doesn’t have one. As Colossal point out, the species has been extinct for thousands of years, and a lot of environmental change has happened between then and now. Dire wolves went extinct at the close of the last geological epoch, the Pleistocene, and part of the reason why we give certain time periods different names is that Earth is fundamentally different on one side of the boundary vs the other. The big changes that happened at the boundary between the Pleistocene and Holocene include the extinction of many megafauna, like dire wolves, sabre-toothed cats, and mammoths. One of the likely reasons why dire wolves went extinct is that the large herbivores they preyed on went extinct, and that’s as true now as it was before.
Colossal says, “The dire wolf’s role in their ecosystem was undeniably crucial. Their absence as apex predators formed a noticeable void—leaving a disruption in the natural order in its wake.” Ecosystems are intricate webs of species and abiotic interactions, and it is true that the extinction of Pleistocene megafauna left a lot of holes in ecosystems, especially in the Americas. Studies in North America show that modern communities are less diverse and occupy a smaller proportion of ecological space than their Pleistocene equivalents. But it is naive to suggest that simply plopping an extinct predator, or even multiple species, into modern ecosystems will make everything like the Pleistocene again. The 11,700 years since the Pleistocene ended may be a blink in geological time, but it is still plenty of time for plants, animals, and ecosystems to evolve. The species that survived have changed in morphology and body size, and not always in the same ways. Species are found in very different places now than they were back then. The climate has changed, too, and not just because of the rapid human-caused climate change that’s happening now. Temperatures over the last few thousand years have been close to what they were about 125,000 years ago during the previous interglacial - about 50,000 years earlier than the older of the two wolves that Colossal sampled genomes from.
One of the big issues with Colossal’s presentation of information as a whole is that they often blur the lines between which facts are about grey wolves and which are about dire wolves. They have a lot of “myth vs fact” sections that, while good for dispelling some persistent myths about grey wolves, are called “dire wolf myth[s]” on their website. They point out ultraviolet vision, but there’s no indication about if they’ve done any studies on the UV vision of canids themselves. Much of the information they display about “dire wolf” behaviour and ecology seems to be simply grey wolf biology, as far as I can tell. We can’t, at this time, tell how dire wolves would have behaved in the past. Without knowing all the genes they modified (we know a couple), we don’t know if any of the life history information, like gestation or age at weaning, for the three wolves they have created is based in actual dire wolf genetics. We likely wouldn’t even know that if we did know all of the modified genes, because the genetic control of complex traits, like those involved in life history, is incredibly difficult to understand.
When they do talk about actual dire wolf science, mistakes and inconsistencies creep in. Take this screenshot, for example:
There’s an oopsie where they clearly meant to edit the card that says “Family Canidae” to say “Genus Aenocyon”, but instead made one that says “Family: Genus.” Okay, whatever.
But they have the dire wolf’s status as “Extinct” with the logo of the International Union for the Conservation of Nature (IUCN) beside it. The IUCN is the organization that maintains the Red List of Threatened Species, which works with scientists globally to classify species into categories like Endangered, Vulnerable, or Least Concern. However, the dire wolf is not listed on the IUCN red list, likely because it’s been extinct for so long that there’s no point in listing it. One of the oldest extinctions documented on the Red list is that of the giant fosa, Cryptoprocta spelea, which went extinct during the Holocene, sometime within the last 2,000 years.
The range map is also…interesting…because it’s definitely not correct. The most obvious to me, of course, is that it doesn’t include the Canadian specimen I worked on with my colleagues Talia Lowi-Merri, Kevin Seymour, and David Evans. But it also doesn’t include any of the South American specimens, even though Colossal elsewhere acknowledges that the species was found in both of these places. And it conflicts with their own other range map:
This range map is its own bundle of fun. I don’t know why the minimum specimen age for what appears to be the Medicine Hat specimen is coloured orange, because we clearly state that the specimen is probably younger than ~43,000 years old. I also don’t know anything about the BC and Saskatchewan specimens indicated here. It’s possible those are undescribed specimens.
They discuss how “fossils have been discovered in the forests of Canada.” I’m sure this could refer to the secret BC/Saskatchewan specimens, but if they’re talking about the Alberta specimen I can assure you that it is not currently in a forest, nor do I think it used to be (although I haven’t tested that).
Something cool is that they claim to have found is evidence of hybridization between the dire wolf lineage and both the Canis/Cuon ancestral lineage and an ancient canid lineage. I don’t know enough about hybridization studies to tell how you would (or if you can) identify hybrid events from an ancient lineage that doesn’t exist anymore and has no DNA to sample. Perhaps there are DNA sequences showing a mix of Speothos/Chrysocyon and Vulpes markers? But then, how do you know that wasn’t just ancestral traits that have just been further evolved in Speothos/Chrysocyon in the intervening years? I would love to chat with someone to help me wrap my head around this.
Very strangely, they have a whole section on their site about the Rancho La Brea Tar Pits (RLB) in Los Angeles, California, but they didn’t sample any specimens from RLB. It’s not strange that they didn’t sample from there - as amazing as RLB is, the asphalt that gives the tar pits their name makes for poor DNA extraction - but it’s a little weird to me that they put so much emphasis on it.
Some of this seems to be to implicate humans in the dire wolf’s extinction. The webpage states that “it seems as though dental fracture became a more common occurrence around the introduction of early man.” However, multiple periods of increased tooth breakage have been identified in dire wolves from RLB, and the periods identified in O’Keefe et al. (2014) are earlier than the earliest evidence for human presence in the area we now call Los Angeles (notwithstanding a pretty hotly contested archaeological site in San Diego that researchers suggested was 130,0000 years old).
Colossal has several videos, including “The Making of the Colossal Dire Wolves,” which is quite flashy with uplifting music - marketing at its finest. Early in the video, Beth Shapiro talks about needing a solution to the biodiversity crisis. This is true, we do need a solution - but de-extinction isn’t one of them. Co-founder Ben Lamm says their “goal is to make extinction a thing of the past,” contradicting what we just heard Shapiro say, which is that extinction is a natural process of life on Earth. Now, I don’t mean to imply that humans aren’t causing vastly more extinctions than would happen naturally - we are! - but ceasing all extinctions is not sound scientific practice, it’s playing god.
Colossal states a need for “wolf restoration and conservation,” but also discuss “working together to reintroduce extinct species and recalibrate nature’s natural balance.” I doubt these two things can happen at the same time. In the making of video, Lamm states that predators are an integral part of ecosystems and cites the example of wolf reintroductions to Yellowstone National Park causing cascading effects that allowed for better environments for species such as beavers and songbirds. This is probably the most-cited example of top-down impacts on an ecosystem, and it’s super cool! Wolves are awesome! Which begs the question… why not use all this capital to, ya know, preserve populations of wolves without distracting the public with claims of reviving extinct species? Dire wolves would surely compete with other canid species, just as grey wolves often suppress the numbers of coyotes. Colossal claims to want to help the critically endangered red wolf, but if they “reintroduce” dire wolves, will the red wolves survive? I would hazard to guess that they won’t.
So what did they do, exactly?
Colossal describes the process in seven steps on the “Science in Action” part of their dire wolf page, so let’s use that as a framework for discussing it.
Step 1. Reference genomes
One of the challenges when we started working on the dire wolf was that we knew of only two dire wolf specimens in the world that contain some useable DNA. (Sven Bocklandt)
I feel like this is a bit misleading. Perri et al. (2021) were able to sequence DNA from five dire wolf specimens, including complete mitochondral sequences and low coverage nuclear genome sequences. This is expanded on elsewhere on the site, where they state that “Non-Colossal Scientists sequenced 46 samples. Only 2 had viable specimens with .25 coverage of genome.” Coverage here appears to be depth of coverage (as opposed to breadth of coverage), which is theoretically \(ReadLength * NumberOfReads / GenomeLength\). A depth of coverage of 0.25 is very low, but not necessarily unexpected for fossil material; depth of coverage for the modern samples in Perri et al. were between 5.6 - 56.0. But Perri et al. don’t report a depth of coverage of 0.25, they report a depth of coverage of 0.01 - 0.23. Elsewhere, Colossal reports that that (presumably) Perri et al. sequenced one quarter of the genome (probably meaning 0.25x coverage) and 0.1% of the genome (actually accurate). Of the five Perri et al. sampled, two specimens (DireSP and DireGB in Perri et al.) yielded over 28 million bases. For reference, the wolf genome has over two billion bases.
Colossal used two of these previously-sampled specimens for their genome work, and it appears that they are the same as DireSP and DireGB in Perri et al. (2021), which would make them VP 1737 (found at Sheridan Pit, Ohio and housed at the Cincinnati Museum Center) and IMNH 48001/52 (found at Gigantobison Bay, Idaho and housed at the Idaho Museum of Natural History). This seems to align with what they have said elsewhere about what the specimens were like. Anyway, it is clear that there are more specimens out there with useable DNA, albeit much smaller fragments than these two.
Colossal gives two different numbers for what depth of coverage they were able to get for their dire wolf specimens. In two places, they claim to have recovered “over 500x more dire wolf DNA,” but elsewhere they say they only recovered 55x more DNA, and in another spot they say they “generated over seventy times more data on the dire wolf genome than was previously achieved.” Assuming this is referring to breadth of coverage, Colossal produced dire wolf genomes with somewhere between 13.5x - 125x coverage. If we’re talking number of bases, Colossal sequenced 154 million to 1.4 billion bases. Maybe they’re referring to a bit of both and they got a 13.5x depth of coverage and a total sequence alignment of 1.4 billion bases. That would be pretty cool, but we need to keep in mind that that would still only be about half of the expected genome size for a dire wolf (assuming I’m understanding everything correctly, and it’s possible I’m not). So even if we CRISPR’d all those genes into a grey wolf embryo, we’d still have something that’s at best half dire wolf.
They also sequenced new grey wolf genomes, targeting indivdiuals with no hybrid genes from domestic dogs. They say they looked at two dozen grey wolves in wolf sanctuaries around the US and found four in which they couldn’t detect dog ancestry. Evidently they settled on a donor named Jackson. He’s pretty cute.
Step 2. Sampling for cloning
Their method for cloning relies on extracting DNA from a type of cell found within blood called an endothelial progenitor cell, or EPC, which plays a role in the regeneration of endothelial tissue in blood vessels. This is supposed to skirt around an issue with regular blood cells that would lead to an animal without a functional immune system. The technology is outside my expertise, so I don’t have anything meaningful to say there. However, it’s worth noting that a part of their site says they’re targeting pericytes, which are a different kind of cell that wraps around endothelial cells.
One of the major upsides that Colossal discusses about this method is that a blood draw is far less invasive than some traditional methods for getting material for cloning, which include tissue biopsies and ear clippings. A blood draw is certainly preferable, but they describe it as routine, and while I will say that that’s true for domesticated or zoo animals, a wild animal still needs to be put under anaesthesia for a blood draw, and anaesthesia always comes with risks.
Step 3. Aligning and comparing genomes
Colossal aligned the grey wolf and dire wolf genomes using a proprietary software made by a company spun off of Colossal, called Form Bio. This is not software that I had heard of before this, but a quick look on Form Bio’s website shows that all the testimonials on the home page are by corporate scientists, two of which are with Colossal. The software leans into AI, which these days gives me a bit of an ick factor, but since I am not a geneticist I can’t comment on whether it’s useful AI or something more akin to ChatGPT.
Step 4. Identifying genomic differences
If you want to gene edit a grey wolf to be more like a dire wolf, you have to know what the differences are. Unfortunately, as I discussed above, we have at most half of a dire wolf genome to go off of. Even Beth Shapiro agrees that this is an unsolved problem:
Discovering which parts of the genome make an animal look and act like that animal is one of the hardest problems in biology. It is unsolved, and something that we have to solve if we want to know which parts of the genome we want to edit to make the grey wolf more like a dire wolf.
So we don’t know what makes an animal look and act like a particular species, but also we’ve definitely de-extincted the dire wolf. Sure, Jan.
I don’t see any information on how many differences there were between the two species, what proportion of base pairs were different, anything like that. The only thing they do specifically mention is differences between dire wolves and grey wolves in the LCORL gene, fully known as ligand dependent nuclear receptor compressor like. This gene has been implicated in the size differences within a large number of mammal species, including humans and dogs. Specifically, a mutation in LCORL causes large body size in some dog breeds, like the great dane.
What’s kind of worrying is that Colossal states that “dire wolves had several deleterious changes in their own version of LCORL” (emphasis mine). Deleterious changes mean that they have a negative effect on the animal. An yet, they confirm that this is one of the genes they modified in grey wolf cells, without discussing if or how they made this change non-deleterious.
The only other information we get from Colossal on what genetic differences they found between grey and dire wolves is the following quote:
Our analysis of the dire wolf genome revealed that they were stunning, with likely light, nearly-white coats, sturdy legs and the unique craniofacial features of a true American superwolf.
We learn a bit more through reporting by New Scientist:
Shapiro could not tell New Scientist how many differences there are but said the two species share 99.5 per cent of their DNA. Since the grey wolf genome is around 2.4 billion base pairs long, that still leaves room for millions of base-pairs of differences.
As we’ll see below, the Colossal team did not make anywhere close to millions of changes in their supposed “dire wolves”…
Step 5. Genome editing
The actual genome eiting was done using CRISPR, a technology that allows for precise changes to an organism’s genetic makeup. You might remember CRISPR from the 2018 story of gene-editied human babies, for which the scientist went to jail. But CRISPR has been used for a ton of purposes, including flipping the on/off switch for genes without actually changing the sequence itself.
Colossal says they have introduced 20 genetic variations across 14 loci into their grey wolf donor genome, but they haven’t yet published a list of those changes. Those 20 changes are probably 20 sequences, so the number of base pairs that were changed is going to be higher, but certainly less than the millions it sounds like the team found.
We know from their website that one of the changes they made is to LCORL, and the others are discussed in the interview Beth Shapiro did with New Scientist. Five of the 20 changes (so, 20%) are actually based on grey wolf genetics, and are mutations known to produce light coats. The other 15 changes come from the dire wolf genome and include the changes to LCORL as well as other changes that affect musculature and ear shape.
In fact, five of those 20 changes are based on mutations known to produce light coats in grey wolves, Shapiro told New Scientist. The remaining 15 are based on the dire wolf genome directly and are intended to alter the animals’ size, musculature and ear shape.
Colossal has talked about the wonder of hearing a dire wolf howl for the first time, and has made claims about how this will help us better understand dire wolf behaviour, but it doesn’t sound like they made any genetic changes that are known to affect these traits. So really we’re getting bigger, beefier grey wolves with grey wolf behaviour and grey wolf vocalizations.
Step 6. Creating an embryo
The process they used for transferring their gene-edited cell nucleus into a donor egg is called somatic cell nuclear transfer. This is the same process that was used to clone Dolly the sheep in 1996, making her the first cloned mammal. This is another area where I’m too out of my depth to make any comments, but as far as I know this is pretty standard.
Step 7. Gestation, birth, and the first six months
The now-gene-edited egg has to go somewhere to gestate, and in Colossal’s case they used large domestic dogs. I’m not a particularly huge fan of the fact that we never get to see these dogs or know much about them. Where did they come from? What happens to them after their “task” is done? I wonder if it is a way to distance from the ick factor of a dog having to birth something so different from itself?
Colossal says the wolves gestated for 65 days and were born by scheduled c-section. I can’t help but wonder what the gestation period of a dire wolf actually would be. Given their larger body size, it should be longer than a grey wolf (with a mean of ~63 days), but we don’t know by how much.
The company maintains that they are super invested in making sure the wolf pups are able to be brought up “naturally” by their surrogate mother, or another lactating female. But, the surrogate for the first two male pups, Romulus and Remus (we never hear anything about the birth of the younger female, Khaleesi) only suckled them for a few days. According to Colossal, the dog was becoming “too attentive” and disrupting regular sleeping and eating schedules. I have no idea what a normal amount of attentiveness is, so would love to hear from some wolf and dog behaviour experts on this. What I do know is that hand-reared domestic dog puppies do have a tendency to become more aggressive, anxious, and socially inept. I hope that was considered when making this decision.
I should note here that Colossal has a husbandry guide on their website, but I haven’t had the chance to read it in detail yet. I will at some point.
The pups were weaned at 8 weeks, 11 days later than the mean of ~45 days for grey wolves, but around the time that domestic dog puppies are weaned. I’m getting data for the grey wolf from the PanTHERIA database, which is a really helpful database; but it should be noted that there is variability in all animals.
The pups are now kept on a 2,000 acre, or ~8 km2 plot of land somewhere in the northern US. That is certainly more than most captive animals get, but grey wolves naturally have much bigger home ranges than this, often hundreds of square kilometres.
One thing that concerns me is the period of socialization. As far as I know, Colossal has kept the older Romulus and Remus together with no other animals. In the wild, a wolf would learn how to be a wolf from other wolves. In captivity, I have seen domestic dogs be used to help young carnivores get their bearings. I don’t know if any of that critically important social learning was allowed to take place here, and if it hasn’t, the wolves will suffer for it.
Colossal has been certified by American Humane, and even boasts a quote from their CEO on their webpage. I am glad that an animal welfare organization feels that they are taking good care of their animals, but I am frankly a bit shocked and concerned that they would certify them. I don’t think knowingly creating animals that cannot, by definition, live with their own kind, strikes me as completely inethical.
What’s next?
Beth Shapiro says that it will be at least a year before we know if Colossal has successfully made the changes they wanted to make. That makes sense, since the wolves aren’t fully grown yet. But what if these animals are not considered a success? What if there are unintended consequences to the wolves’ health? Colossal has done this experiment with a large, relatively long-lived, social animal. I don’t see enough evidence to say they have adequately considered these animals’ quality of life.
What is clear is that Colossal doesn’t plan to stop at the three non-dire wolves they already have. Matt James is shown in one of their videos saying:
When we edit different lines of genetically diverse grey wolves, we’ll create a set of genetically diverse dire wolves, with similar edits, but the basis of their diversity already existed within the grey wolf.
This seems to be a somewhat word-salady way of saying that they plan to make genetically modified wolves using not just Jackson as a donor, but many more grey wolves. Colossal says that they don’t plan to breed Romulus, Remus, and Khaleesi, but the only reason they would want genetically diverse “dire wolves” is if they plan on them breeding.
And they have support from the current US government. Colossal displays a worrying quote from Doug Burgum, US Secretary of the Interior on their website, which says that “it has been innovation—not regulation—that has spawned American greatness.” It sure sounds like there will be no governmental control on what companies like Colossal will do to our ecosystems.
Colossal thinks that these animals should be in the wild, but that’s a horrendously bad idea. As I said toward the beginning of this article, we would run the risk of these facsimile dire wolves out competing the beautiful, wonderful grey wolves and red wolves that some people are trying so hard to preserve. Wolves have already suffered so much persecution, loss of habitat, and loss of prey by human hands. Do we really need to add this to the mix?
Colossal claims that this “officially marks the end of the rampant species eradication which plagues our planet.” I call bullshit. Colossal is not addressing the causes of species extinctions, they are merely dangling a shiny toy to distract us from the real problems at hand.