Bringing Extinct Species Back to Life | Dr. Beth Shapiro

Beth, I’ve followed your work for a long time. You’re an evolutionary biologist helping lead Colossal, where mammoths, dodos, thylacines, and dire wolves are not just movie premises. I want the science, ethics, and conservation upside—but start ridiculously basic: what counts as a species? Are my bulldog mutt and a Chihuahua really the same thing?

Biology does not care about our labels. Species are human boxes that help us tell stories and know what we mean. The familiar definition is organisms that reproduce and have fertile offspring, but that is only one useful definition. You can classify by DNA, geography, or conservation need. Taxonomy can be wonderfully silly: the plains bison is Bison bison bison.

Neanderthals make that messier. We learned a neat ladder of human evolution, but ancient DNA says modern humans met Neanderthals after leaving Africa and had children with them. Were they another species, or do the words stop helping right when the story gets interesting?

My husband helped assemble the Neanderthal genome, and the evidence is clear: several human-like lineages lived at once, including Neanderthals and Denisovans. Modern humans interbred with Neanderthals, and many people carry roughly two to five percent Neanderthal ancestry. Across people worldwide, we can reconstruct most of the genome from surviving fragments. Some variants connect to immunity, disease risk, pain, or metabolism—but “more Neanderthal” is not a vigor score.

Evolution is slow, contingent, and not a march from four legs to two. Isolated populations change through selection and chance; when they meet again, hybrids may be impossible, viable, or viable only in one environment. Polar and brown bears can cross, but polar-bear ancestry persists in brown bears, not the reverse. A partly brown bear may not be white enough to hunt seals.

Why mammoths instead of dinosaurs, Florida panthers, or every creature we are losing now? People hear de-extinction and picture Jurassic Park: make one animal, fling it into Alaska, and wait for the fences to fail.

Dinosaurs are simply out: their DNA is gone. DNA breaks down after death, though cold, rapid burial can preserve mammoth DNA for around one to two million years. We need preserved samples, ideally many, and we need to understand why an animal vanished so we do not bring it back only to lose it again.

Choice is technical, ethical, ecological, and social. Big herbivores move nutrients, disturb soil, spread seeds, and change vegetation. At Pleistocene Park, animals clearing winter snow expose ground and support a more varied plant landscape. The point is not that a mammoth fixes every ecosystem; it is whether restoring an interaction can make a particular place tougher under pressure.

I pushed for a bird program because birds are among the most endangered animals and cannot be cloned like mammals. The dodo gives people an immediate feeling—yes, its ridiculous beak helps—and awe matters. Mammoths, dodos, and thylacines can bring attention and investment into tools that help species still here.

Jurassic Park was not a documentary, bless it. We are not filling a mammoth genome’s gaps with random DNA. Asian elephants are their closest living relatives. We compare many mammoths and elephants, focus on changes tied to mammoth traits, then engineer an elephant able to live safely in present and future cold environments.

The dire-wolf publicity got people arguing over the name. But the question is not just a strict sequence threshold: what did you change, and what does the animal look like now?

Our dire wolves have twenty selected edits, based on fossil genomes, in a gray-wolf background. They are larger, more muscular, long-haired, and light-coated. Safety leads every choice: fossil light-coat variants sit near changes that can cause blindness or deafness in living dogs. We used different, known-safe edits that produce the same appearance. There are three: Romulus and Remus, nearly two-year-old brothers, and Khaleesi, a younger female. We are not releasing them; they would compete with gray wolves.

Conservation is not the same as leaving nature untouched. We already choose which animals get protected, vaccinated, fed, or moved. The question is whether intervention improves resilience. In Tasmania, a missing thylacine may matter because a top predator could have removed sick devils before facial cancer spread so disastrously.

And de-extinction is not instead of helping living species. Northern quolls die after eating invasive cane toads. Elsewhere, animals have a one-letter genetic difference that tolerates the toxin. We have made that change in quolls and are testing it. The same comparative genomics, editing, and verification can keep a species from disappearing.

That is the trust problem. People hear gene drives, edited animals, or AI modeling ecosystems and think scientists are doing things because they can. We need different views in the room, transparent risk analysis, and a real answer to who gets to decide—not one expert declaring the future.

I agree: care, local participation, and discussion must come before an animal exists to release. Māori-led Moa work in New Zealand puts long-term stewardship with the people who will live with it. But doing nothing is still a choice. Ecosystems are changing too fast for natural selection alone to rescue every species, and refusing every tool means accepting a less diverse future.

Black-footed ferrets show why multiple tools matter. Prairie-dog control devastated their food supply; they were rediscovered after a family dog killed one in Wyoming. Captive breeding restored numbers but left little diversity. Cloning preserved tissue introduced lost variation, and future editing might address the plague that kills them in the wild. Genetic rescue is not one magic trick.

Humans are the deepest water. We already select embryos in IVF, but disease avoidance is very different from chasing intelligence or height. The scary part is inequality, coercion, and deciding what another generation ought to be.

We have always shaped life by choosing mates, domesticating wolves, turning teosinte into corn, and moving species around. Editing people feels different because we do not want to assign a child a narrow role before they choose their own life. Yet a bespoke CRISPR treatment reportedly cured baby KJ’s urea-cycle disorder. These tools can be frightening and spectacular at the same time.

Ancient DNA lets me read stories from whole ecosystems that are gone: warming after an ice age, humans arriving, species adapting or vanishing. Mammoths make us feel small in a way we can appreciate. If people are arguing, learning, getting mad, getting excited, and feeling they have agency in the world they inherit, that is a good place to start.