Colossal Biosciences is a biotechnology company attempting something no lab has done before: using gene editing, stem cell engineering, and advanced reproductive technologies to bring back extinct species, starting with the woolly mammoth, the thylacine (Tasmanian tiger), and the dodo. Founded in 2021 with high-profile backing and hundreds of millions in venture funding, Colossal has become the most visible player in de-extinction, a field that sits at the intersection of paleogenomics, synthetic biology, and conservation. Whether their ambitious pipeline produces living animals or merely advances the underlying science is an open question, but the technical and ethical terrain they are navigating tells us a great deal about where biology is headed.
What Colossal Is Actually Trying to Do
Colossal’s flagship project aims to produce a cold-adapted elephant that carries key woolly mammoth traits: dense hair, smaller ears, a thick fat layer, and hemoglobin that functions efficiently in freezing temperatures. The company’s approach involves editing Asian elephant cells using CRISPR to introduce mammoth-specific genetic variants, then using those engineered cells to create embryos that could eventually be gestated and born. The end product would not be a pure woolly mammoth. It would be a hybrid, an Asian elephant chassis with mammoth features layered in. Colossal describes these animals as “functional proxies” for the extinct species.
The thylacine project follows a similar logic but with a different surrogate lineage. Researchers at the TIGRR Lab, with which Colossal collaborates, are working to modify the genome of the fat-tailed dunnart, a small marsupial, to introduce thylacine DNA via CRISPR. The marsupial reproductive system complicates surrogate pregnancy in ways that differ sharply from placental mammals.1PubMed Central. Molecular Paleontology Meets Drug Discovery: The Case for De-extinct Antimicrobials The dodo project, announced later, targets the Nicobar pigeon as the closest living relative, though avian genome editing poses its own set of hurdles because bird eggs are difficult to manipulate at the single-cell stage.
Reading the Genomes of the Dead
Before you can edit mammoth genes into a living elephant cell, you need to know what those genes looked like. Ancient DNA degrades over time, accumulating chemical damage that can masquerade as real genetic variation. A simulation study modeling DNA damage from a roughly 120-year-old specimen found that even at 30× sequencing depth, degraded samples generated over 1.4 million false genetic variants. Only at extremely high depth, around 110× or more, did false positives drop to negligible levels.2PubMed Central. Mapping the Genomic Limits of De-Extinction in the Face of Ancient DNA Degradation Most paleogenomics studies sequence specimens to far less than 30×, and even 10× is unusual. For woolly mammoths, whose remains are thousands to hundreds of thousands of years old, the damage is far worse than in a century-old sample.
This means Colossal cannot simply read a mammoth genome and trust every variant it finds. The team relies on comparing multiple mammoth specimens to each other and to living elephants, filtering out damage-induced artifacts and zeroing in on variants that are consistently present across individuals. The quality of the reference genome matters enormously. If a variant shows up in one degraded mammoth specimen but not in 22 others, it is likely noise. If it appears fixed across all sampled mammoths and absent in elephants, it is a candidate for a real functional change.
Which Genes Make a Mammoth
Comparative genomics studies have identified thousands of protein-altering mutations unique to woolly mammoths. An analysis of 23 mammoth genomes, including one of the oldest known specimens at roughly 700,000 years old, identified about 3,100 genes carrying mammoth-specific mutations predicted to affect protein function. Many of these genes are associated with hair and skin development, fat storage and lipid metabolism, immune function, temperature sensation, circadian rhythms, and DNA repair.3Current Biology. Multiple Genomes Provide Insights into Genomic Changes Associated with Pleistocene Adaptations in Woolly Mammoth The breadth of these changes is striking: the woolly mammoth had already acquired most of its signature adaptations early in its evolutionary history, suggesting strong and rapid selection for cold tolerance.
One of the better-understood examples involves a gene called TRPV3, which encodes a channel involved in temperature sensing and hair growth. Researchers resurrected the mammoth version of this gene in the lab and showed that a single amino acid substitution strongly altered the channel’s sensitivity to temperature.4PubMed. Elephantid Genomes Reveal the Molecular Bases of Woolly Mammoth Adaptations to the Arctic That kind of functional validation is rare. For most of the 3,100-plus candidate genes, we know the mammoth carried a different version, but we do not know exactly what it did. Colossal’s task is to decide which of these thousands of edits to prioritize, since editing all of them into an elephant cell is not currently feasible.
Earlier comparative work confirmed the general pattern: mammoth-specific genetic changes cluster in functions related to lipid metabolism, circadian biology, skin and hair physiology, and immune defense.5PubMed Central. Evolutionary adaptation revealed by comparative genome analysis of woolly mammoths and elephants The challenge is not just identifying variants but understanding how they interact. Traits like dense fur or cold-adapted fat storage likely depend on networks of genes working together, not single switches.
From Edited Cells to Living Animals
Colossal’s gene editing work begins with induced pluripotent stem cells (iPSCs), which are adult cells reprogrammed to behave like embryonic stem cells, capable of becoming any cell type. The company has reported creating elephant iPSCs, a significant milestone given that reprogramming has been demonstrated in relatively few non-model species. The core molecular machinery for reprogramming appears deeply conserved across vertebrates, suggesting it could work in a wide range of animals, though results in distant lineages remain uncertain.6PubMed Central. Induced Pluripotent Stem Cells in Non-Model Species: Applications and Challenges
Once mammoth-specific edits are introduced into elephant iPSCs, the next step is turning those cells into an embryo. The most discussed method is somatic cell nuclear transfer (SCNT), the same technique used to clone Dolly the sheep. A donor cell nucleus is fused with an egg cell stripped of its own nucleus, triggering embryo development. For de-extinction, this would be interspecies SCNT, since the egg would come from an Asian elephant or possibly a closely related species. Because biological remains degrade after death, cloning long-dead organisms directly from preserved tissue is not feasible, which is why Colossal works through the iPSC-then-editing route rather than attempting to clone from ancient mammoth cells.7Functional Ecology. Pathways to de‐extinction: how close can we get to resurrection of an extinct species?
Getting an interspecies embryo to develop inside a surrogate mother is one of the hardest parts. Interspecies gestation has succeeded in several wildlife species, including endangered gaur, mouflon, African wildcats, sand cats, and gray wolves. But success rates tend to be low, and potential causes of failure include immunological rejection, problems with the placenta, and mismatches in gestation length or fetal growth rates. The most famous cautionary tale is the bucardo, a subspecies of Spanish ibex: a clone was born alive in 2003 but died within minutes due to lung defects potentially linked to interspecies gestational incompatibilities.8Journal of Heredity. De-extinction technology and its application to conservation
Elephant pregnancies last about 22 months, the longest of any land animal. The logistics of producing elephant eggs, performing nuclear transfer, and supporting a nearly two-year pregnancy in a surrogate are formidable. Colossal has indicated interest in developing artificial womb technology as an alternative, which would bypass the surrogate problem entirely.
The Artificial Womb Question
Artificial womb technology, sometimes called ex vivo uterine environment therapy, aims to replicate the conditions of the uterus outside the body. Current systems are designed to support extremely premature human neonates by substituting core placental functions like gas exchange and nutrition while keeping the fetus in a warm fluid environment with strictly umbilical access.9PubMed Central. Artificial Womb Technology: A Systematic Review of Preclinical Evidence and Implications for Neonatal Viability and Intensive Care Existing models have established partial extrauterine environments for premature fetuses, enabled early embryo culture in the lab, and promoted development of uterine organoids and engineered endometrial tissue.10Chemical Engineering Journal. Artificial womb technology: Progress, challenges, and future directions toward extrauterine gestation
There is a vast gap between supporting a late-stage premature lamb for a few weeks and gestating an elephant-mammoth hybrid from embryo to full term. No artificial womb has ever sustained a mammalian pregnancy from its earliest stages through birth. The technology is promising for neonatal medicine, but applying it to de-extinction would require leaps in scale and duration that are not close to being demonstrated. For Colossal, artificial wombs represent a long-term moonshot within an already moonshot-level project.
A Hidden Obstacle Inside Every Cell
Even if the nuclear genome is edited perfectly, a second genome sits inside every cell: mitochondrial DNA. In interspecies nuclear transfer, the mitochondria come from the egg donor species, not from the edited nucleus. That means a mammoth-edited elephant nucleus would be paired with Asian elephant mitochondria. This mismatch can cause real problems. Incompatibility between mitochondrial and nuclear genomes has been linked to failures in interspecies cloning, including issues with embryonic genome activation and abnormal development.11PubMed Central. The perspective of the incompatible of nucleus and mitochondria in interspecies somatic cell nuclear transfer for endangered species
For the mammoth project, the nuclear-mitochondrial gap may be manageable because Asian elephants and woolly mammoths are closely related, having diverged only a few million years ago. For more distantly related species pairs, the incompatibility grows worse. The thylacine and the fat-tailed dunnart, for instance, are separated by tens of millions of years of evolution, making cytonuclear mismatches a more serious concern. Colossal has not publicly detailed how it plans to address this for each project, and the problem does not have a clean solution yet.
What Would a De-Extinct Animal Actually Be For
Colossal frames its mammoth project partly as an ecological intervention. The idea draws on a hypothesis that large herbivores once maintained Arctic grasslands by trampling shrubs, compacting snow, and recycling nutrients. Today’s tundra is dominated by mosses and shrubs, and the underlying permafrost is thawing, releasing stored carbon. The proposal is that reintroducing large grazers could help revert some of this landscape to grassland, potentially slowing permafrost thaw and reducing carbon emissions, especially in carbon-rich permafrost regions.12PubMed Central. Pleistocene Arctic megafaunal ecological engineering as a natural climate solution?
The ecological logic is real but speculative at the scale Colossal envisions. Pleistocene Park, an experimental reserve in Siberia, has been testing the grassland-restoration hypothesis with existing large herbivores like bison, horses, and muskoxen for decades. Results suggest that herbivore activity does change vegetation and snow dynamics, but whether a handful of cold-adapted elephant hybrids would meaningfully affect permafrost across the vast Arctic is another matter. The ecological argument provides a compelling narrative, but critics point out that the same funds could support proven conservation strategies for species that still exist.
The Genetic Diversity Problem
A de-extinct population born from gene editing would start with extremely low genetic diversity, potentially from a single edited cell line. Small founder populations face inbreeding depression, where harmful genetic variants accumulate as closely related individuals reproduce. Research on the northern white rhino, which has been reduced to two living females, offers a useful parallel. Simulations of a population restored from biobanked cells found that repeatedly reintroducing new founders from banked genetic material could keep fitness costs from genetic load below those already tolerated by southern white rhinos, a viable wild population. Without such repeated reintroductions, the restored population would need rapid growth rates exceeding 20 to 30 percent per generation to maintain comparable fitness.13PubMed Central. Genetic load and viability of a future restored northern white rhino population
For a de-extinct mammoth proxy, the situation is even more constrained. There are no living mammoths from which to bank cells. All genetic diversity must be reconstructed from ancient DNA or engineered synthetically. Colossal could, in theory, create multiple distinct cell lines carrying different combinations of mammoth variants, but each line would require its own round of extensive editing and validation. Building a genetically diverse founding population would be extraordinarily expensive and slow.
Where De-Extinction Tools Help Living Species
Perhaps the most tangible near-term value of Colossal’s work lies not in reviving mammoths but in developing technologies that benefit endangered species that are still alive. The tools being refined for de-extinction, including iPSC reprogramming, interspecies cloning, and precision gene editing, are directly applicable to conservation biology. Each of these technologies can be used for what some researchers call “de-endangerment,” offering new approaches to preserving biodiversity in species teetering on the edge.14Journal of Heredity. De-extinction technology and its application to conservation
The black-footed ferret illustrates this well. Once thought extinct, the species was rediscovered in the 1980s but carries severely depleted genetic variation because the entire living population descends from just seven individuals. Genome resource banks hold genetic material from ferrets not represented in the current lineage. Researchers have proposed using interspecies cloning to reintroduce that lost diversity, a strategy that could serve as a model for other genetically impoverished species.15PubMed Central. A Road Map for 21st Century Genetic Restoration: Gene Pool Enrichment of the Black-Footed Ferret In 2020, a black-footed ferret named Elizabeth Ann was cloned from cells frozen in the 1980s, becoming the first cloned U.S. endangered species. That work was done by a different organization, but the underlying techniques overlap substantially with what Colossal is developing.
Ethical Fault Lines
De-extinction sits in contested ethical territory. Proponents argue it could restore lost ecological functions, inspire public interest in conservation, and push biotechnology forward. Critics raise several concerns. Animal welfare is a significant one: creating an animal through interspecies cloning and extensive genetic modification carries real risks of suffering, as demonstrated by the high failure rates and developmental abnormalities seen in cloning experiments. The bucardo clone’s death minutes after birth is not an anomaly in the history of SCNT; many cloned animals are stillborn or suffer health problems.
There is also the question of hubris. Some ethicists argue that de-extinction promotes a dangerous confidence in technology’s ability to fix ecological damage, potentially weakening the public urgency around preventing extinctions in the first place.16Biological Conservation. Between hype and hope: De-extinction is a tool, not a panacea for the biodiversity crisis If the public believes extinction is reversible, the political will to protect habitats and living species could erode. Media coverage of de-extinction tends toward spectacle, which amplifies this risk.
Resource allocation is the most pragmatic objection. Conservation biology is chronically underfunded. Hundreds of millions of dollars flowing into de-extinction projects could, in principle, protect vast tracts of habitat or fund captive breeding programs for species on the brink. Colossal counters that its funding comes from private venture capital, not from government conservation budgets, and that the technologies it develops will benefit living species. Whether that argument holds over time depends on whether the spillover benefits materialize at meaningful scale.
Legal Gaps No One Has Filled
If Colossal succeeds in producing a living mammoth proxy, it would enter a world with no clear legal framework for its existence. Nature conservation law in both the United States and the European Union was written to protect existing species and their habitats. A de-extinct animal does not fit neatly into any existing category. Is a cold-adapted elephant hybrid a woolly mammoth for legal purposes? Would it qualify for protection under the Endangered Species Act? Could it be patented as a genetically engineered organism? These questions have been mapped by legal scholars but remain unresolved.17Oxford Academic (Journal of Environmental Law). De-Extinction, Regulation and Nature Conservation
The regulatory ambiguity extends to release into the wild. Introducing a genetically modified large mammal into an Arctic ecosystem would likely trigger environmental review processes, but existing frameworks for GMO regulation were designed for crops and lab organisms, not for megafauna intended to roam freely. Colossal’s projects may force governments to develop entirely new categories of environmental regulation, a process that typically moves far slower than the technology it is trying to govern.
How Marsupial Biology Complicates the Thylacine Project
The thylacine project faces a biological challenge the mammoth project does not: marsupial reproduction. Marsupials give birth to highly undeveloped young that continue growing in a pouch, which means the gestational window inside the uterus is much shorter than in placental mammals, but the postnatal development period is critical and species-specific. The fat-tailed dunnart, the intended surrogate, weighs about 10 to 20 grams. A thylacine weighed up to 30 kilograms. Even if an edited dunnart embryo could be gestated to birth, the massive size difference between surrogate and target species raises questions about whether the pouch environment, milk composition, and developmental cues would support the growth of a much larger animal.
Marsupial iPSC work is also less mature than the equivalent research in placental mammals. The molecular pathways controlling pluripotency are broadly conserved, but the specific factors, culture conditions, and reprogramming timelines needed for marsupial cells are still being worked out. The thylacine project, in some respects, is pushing the boundaries of basic marsupial biology as much as it is pursuing de-extinction.
When Ancient Specimens Run Out of Information
A subtler limitation applies to all de-extinction projects: genomes do not encode behavior. A mammoth proxy might carry genes for dense fur, cold-adapted hemoglobin, and thick subcutaneous fat, but mammoths were also social animals with complex behaviors presumably transmitted through learning within herds. Migration routes, foraging strategies, social hierarchies, and communication patterns cannot be recovered from DNA. A newborn mammoth proxy raised by Asian elephants, or worse, raised without any elephant social group at all, would face a behavioral blank slate.
The same problem applies to the thylacine. Thylacines were apex predators with hunting behaviors, territorial patterns, and social structures that no living marsupial shares closely. The passenger pigeon, another high-profile de-extinction target pursued by a different organization, faces a related challenge: the species was famous for its enormous flocks, a social behavior that shaped its ecology. Reconstructing complex behaviors from genetic data alone remains beyond current science.1PubMed Central. Molecular Paleontology Meets Drug Discovery: The Case for De-extinct Antimicrobials An animal with a mammoth’s physiology but an elephant’s behavior occupies an ecological niche that has never existed.