Bringing Extinct Species Back to Life | Dr. Beth Shapiro

KEYPOINTS

• 01

  'Species' is a human label, not a biological fact

  Shapiro says 'biology doesn't care what species you are.' The biological species concept (Ernst Mayr: can two animals interbreed?), genetic-similarity cutoffs and geographic labels (Florida vs. Texas panther) are all valid tools for different conversations. Her example: Linnaeus named American bison 'buffalo' because Europeans gave that name to any big animal that looked like it would make a good coat.

• 02

  Interbreeding between related lineages is common

  Humans and Neanderthals split about 300,000–500,000 years ago. Brown and polar bears split about half a million years ago and still interbreed wherever their ranges overlap, producing 'grolar' or 'pizzly' bears. Ancient DNA from a polar bear that lived off Alaska more than 100,000 years ago already carried brown-bear ancestry.

• 03

  Everyone carries a different 2–5% of Neanderthal DNA

  The Neanderthal pieces in one person's genome differ from another's. Pooled across living people, they rebuild more than 90%, possibly over 95%, of the Neanderthal genome. So most Neanderthal DNA wasn't harmful. Most of it sits at about 3% frequency and isn't under selection. The interesting signal is a Neanderthal segment that is suddenly common in a population, which suggests it helped.

• 04

  Neanderthal genomes help pinpoint what makes us human

  Comparing humans with chimps means sifting 3–5 million years of change, most of it neutral copying errors; every child has about 100 new mutations compared with its parents. The Neanderthal genome cut the window to 300,000–500,000 years. The roughly 5% of the genome where no living person carries Neanderthal DNA is where the human version was apparently essential.

• 05

  De-extinction means editing a living relative, not Jurassic Park

  No DNA survives from dinosaurs, which died out more than 66 million years ago. The oldest DNA ever recovered is from mammoth bone about 1–2 million years old. Colossal sequences extinct genomes, finds the key differences from the closest living relative, and edits those into the relative's cells. Mammoths are closer to Asian elephants than Asian elephants are to African elephants, and the two genomes are already nearly identical.

• 06

  The dire wolves are gray wolves with 20 targeted edits

  From fossil dire wolf genomes the team identified variants for bigger size, sturdier build and a light coat, then engineered 20 edits into gray wolf cells. For the light coat they used a route already known to be safe instead of copying the dire wolf variant directly. Romulus and Remus are almost two years old and Khaleesi is about 18 months. They're too closely related to breed, so hormones are used to prevent it.

• 07

  A mammoth is a much harder engineering problem

  A woolly coat requires changes not just to hair but to skin: room for more follicles and sebaceous glands. Elephants carry for 22 months and mature at about 14 years, so any edit has to produce a healthy, long-lived animal. Because hundreds of mammoths can't come from hundreds of pregnant Asian elephants, Colossal is developing artificial wombs, for now in mice.

• 08

  Birds can't be cloned, so the dodo drives new tools

  Somatic cell nuclear transfer, the Dolly-the-sheep method used for mammals, doesn't work in birds because their egg cells can't be accessed at the right stage. Birds are among the most endangered animals, so the dodo program is meant to develop bird technologies. It was picked partly for wonder: the 'cool beak' and its Alice in Wonderland fame draw public interest.

• 09

  De-extinction and conservation use the same toolkit

  Shapiro's answer to 'why not save living species instead?' is that Colossal does both. Excitement about mammoths, dodos and thylacines brings in funding and engagement for technology that also protects living species. Humans have been reshaping ecosystems for tens of thousands of years, by hunting large animals in Australia about 50,000 years ago and in Asia about 30,000 years ago, and then through domestication.

• 10

  Gene drives can be built with limits

  Rather than wiping out mosquitoes, Shapiro favors engineering them so they can't carry disease. A gene drive spreads an edit through the wild population, but it can be designed to last only a set number of generations, with appropriate safety measures.