88 | Neil Shubin on Evolution, Genes, and Dramatic Transitions
What this covers
Neil Shubin, a paleontologist and co-discoverer of Tiktaalik Roseae, speaks with Sean Carroll about how evolutionary history actually works—focusing on a counterintuitive pattern: the traits we think define major biological revolutions arrived millions of years before those revolutions happened. Lungs, limbed fins, and feathers did not emerge when fish invaded land or when dinosaurs took to the air. Instead, these structures evolved in earlier contexts, within water or for other purposes, and were later repurposed for new roles. The conversation traces this pattern through fossils, genetics, and development, arguing that [repurposing of pre-existing structures](concept:Repurposing in Evolution) is how evolution typically operates at moments of dramatic change.
Shubin ranges across multiple scales of evidence. He discusses Tiktaalik itself—a genuine mosaic of fish and tetrapod features, with both lungs and gills and a fin containing limb bones. He moves into molecular biology, explaining how a shared genetic toolkit generates the diversity of animal forms: the genes that build fingers in humans build fin rays in fish, with differences arising not from new genes but from genetic switches turning old genes on and off in new combinations. He touches on gene duplication as a major source of evolutionary novelty, the work of Susumu Ono on chromosomal material and genome size, and the domestication of ancient viruses—notably the Arc gene, which Jason Shepherd showed evolved from a viral capsid and now carries genetic material between neurons to form memories. The discussion also covers evolutionary trade-offs in humans, the concept of "missing links" as a misconception rooted in linear thinking rather than the actual branching bush of evolution, and how modern biology has expanded Charles Darwin's framework with genetics while keeping natural selection and common descent at its core.
Shubin argues that major evolutionary transitions (fish to land, origin of flight) don't require wholesale new inventions but rather the repurposing of pre-existing structures, genes, and developmental processes that arose millions of years before the transition itself.
- Lungs, limb-bone-bearing fins, and air sacs existed in fish well before the move to land
- Major appendages across species share a common genetic toolkit, with differences arising from how old genes are switched on and off rather than from new genes
- Even viruses get domesticated and repurposed into functional genes like Arc, which is involved in memory
Major evolutionary traits arose millions of years before the revolutions they are associated with, revealed by transitional fossils and shared genetic structures.
- Lungs did not arise for living on land; they originally evolved in fish living in water, developing as an out-pocket of the gut tube. Many fish have an air sac connected to the gut tube that becomes either lungs or a swim bladder, and lung-bearing fish gulp air at the surface to survive when water oxygen concentration is too low.
“No, they're not. They originally came about in creatures living in water and we've known that since Napoleon's expeditions to Egypt in the early 1800s”
- Pretty much every trait we associate with a great evolutionary revolution—limbs and the invasion of land, lungs and the invasion of land, feathers and the origin of flight, walking on two legs and the origin of humanity—is not actually associated with that revolution but came about millions of years before it.
“Pretty much every trait that we associate with a revolution is not associated with that revolution. It came about millions of years before.”
- Tiktaalik is a genuine mosaic between fish and limbed animals: it had a flat head like a tetrapod, scales and fin rays like a fish, a fin containing a humerus, radius, ulna, and wrist (capable of doing a push-up), plus both lungs and gills—providing real physical evidence of how the water-to-land transition happened.
“So it was a real mosaic between characteristics of fish and limbed animals. And it was an animal that had lungs and gills, had fins with limb bones inside.”
- The genes that build wrists and digits in mice and people are present and active in fish, where they make the terminal strip of tissue that gives rise to fin rays; the difference is a switch in fish that turns those gene-active cells into fin rays whereas in tetrapods they make fingers and toes—revealing a common genetic history of fins and limbs, with differences arising from using old genes in new ways rather than inventing new genes.
“many of the genes that make the wrists and digits of mice and people and birds and so forth, are present in fish. And they're making a terminal strip of tissue in those fins.”
- The motivation for moving onto land was likely the escape from a fish-eat-fish Devonian water world full of 15-foot predators with railroad-spike-sized teeth, into a land environment that had plants, invertebrates, and food sources but few competitors and no predators—so any trait allowing a creature to access land would be favored by natural selection.
“any trait that would allow a creature to avoid the water, get away from the predators and the competitors into the mudflats and to land would probably be favored, because there was a world of opportunity with foodstuffs on land without competitors or predators”
- Finding key transitional fossils is a predictive science: by knowing the evolutionary history of groups, the ages when fossils first appear, and the environments creatures lived in, paleontologists target locations with rocks of the right age, the right type to preserve fossils, and the right accessibility (exposed, not buried or in war zones), putting the odds in their favor rather than searching randomly.
“We didn't go randomly to the Arctic, and we didn't just stay there... we felt we had a very strong prediction that the odds were in our favor if we stuck with it long enough that we would be successful”
All animals share ancient genes and genetic toolkits from common ancestors, with differences arising from reusing old genes in new ways.
- Genes can be experimentally swapped across species and still function: a shark fin gene placed into a mouse works perfectly fine in the mouse limb, and a mouse gene placed into a shark works in its fin, demonstrating a deep genetic connection among all life on the planet.
“I could take a shark gene from the fin and put it in a mouse, some of these genes from a shark and then put them into a mouse. And it does perfectly fine.”
- The same genes that build animal bodies—such as those determining where a wing versus an antenna appears in a fly—are not unique to flies but exist in versions across salamanders, frogs, worms, mice, and people, forming a common genetic toolkit doing similar things in many different creatures, indicating these are ancient genes inherited from a distant common ancestor rather than convergent evolution.
“these genes aren't present only in flies, but versions of these same genes are present in salamanders, frogs, worms, mice and people. That there's a common toolkit.”
- All cells in the body have essentially the same DNA; what differs between a retina cell and a fingertip skin cell is which genes are turned on and off, governed by genetic switches—a tiny sequence next to a protein-coding gene that, when something attaches to it, turns the gene on or off—operating within a dynamic three-dimensional genome that opens and closes and folds to touch these switches.
“the cells in the retina of our eye, compared to the cells in the skin of our fingertip, they have the same DNA inside them, but the DNA in the retina cells is making proteins that builds and keeps a retina functioning... what's different here is which genes are turned on and off”
- The protein-coding portion of the human genome—the genes that contain information to make proteins—is only about 2% of the genome; the rest controls the activity of those genes through regulatory switches, with much remaining mysterious or possibly related to the dynamic three-dimensional geometry of the DNA rather than individual sequence.
“in humans, about almost 10%, about 8% of our genome, of our entire genome are defunct viruses. Viral sequences that no longer function, but they're there. It's like a fossil graveyard”
- Many conceptual tools used in modern evolutionary biology originated in the pre-Darwinian world when scientists sought order and rules underlying the diversity of fixed, divinely-created species; the concept of homology—comparing similar structures (like the same vertebra) across different species—had pre-Darwinian roots, but Darwin gave it a mechanistic, material explanation: these structures are similar because the species share an evolutionary history.
“this notion called homology... the idea that you can compare similar structures in different species... That was a notion that came up in the pre-Darwinian world, but yet, Darwin basically put a mechanistic... A material spin on it”
- Susumu Ono showed using a low-tech method—photographing chromosomes, printing the pictures, cutting them out, and weighing them as a proxy for DNA quantity—that salamanders have about 10 times more chromosomal material than humans, and lilies have enormous amounts, demonstrating that the amount of genetic material in a cell is unrelated to the complexity of the organism.
“the salamanders have 10 times more genes, chromosomal stuff, than humans. And he was one of the first persons using this low-tech technique to show that the amount of genetic material in the nucleus... is unrelated to the complexity of the organism.”
Evolution proceeds through accumulation of small mutations channeled by shared genetic recipes, not sudden macromutations or random chance.
- Mutation is not truly random; it is only random with respect to predicting the future—just because a creature may need to walk on land in 100 million years does not mean it will produce mutations to do so. But with respect to everything else, certain mutations are more common and more likely to be beneficial than others based on what happens in the genome and in development.
“It's only random with respect to predicting the future. Just because a creature may need to walk on land in 100 million years, doesn't mean it's going to come up with mutations to do that.”
- If organisms have the same set of genes functioning in similar ways, it makes similar evolutionary outcomes more likely, so the same evolutionary invention can appear independently in unrelated species (multiples); parallel evolution is not purely random because shared genetic recipes channel the kinds of mutations natural selection can work on—like different bakers with the same cupcake recipe arriving at the same cupcakes.
“if organisms, if creatures have the same set of genes functioning in the same ways, it makes similar outcomes more likely... if organisms have similar genes and they're using those genes in similar ways, then they might produce the same kinds of mutations independently, that natural selection can work on. So it's not random.”
- Goldschmidt's 'hopeful monster' theory—that major evolutionary transitions can happen in one step via a single set of macromutations (e.g., the first bird hatching from a reptile egg)—is not a viable theory because such big mutations are almost exclusively lethal; instead it is the accumulation of small, beneficial mutations that drives evolution, though that accumulation can happen much faster than geological timescales.
“it just turns out that those big mutations are almost exclusively lethal. It's really the small mutations that are most likely to be beneficial, so it's the accumulation of these.”
- Susumu Ono observed that whole sections of striped chromosomes appeared duplicated as if Xeroxed, leading him to propose that gene duplication—duplicating old genes—is a major source of innovation for new genetic material in evolution, a profound discovery confirmed by modern sequencing showing whole gene families containing hundreds of genes related by rounds of historical duplication.
“perhaps one of the major sources of innovation for new genetic material in evolution is gene duplication, is duplicating old genes. So that's a great way of co-opting and repurposing.”
Modern evolutionary biology extends Darwin through genetics and molecular evidence while keeping natural selection and common descent central.
- In science, solving one problem opens up multiple new puzzles, and as we gain answers we get to ask new, more powerful, more precise, or more global questions—the questions never stop, they just change.
“sometimes you actually get to ask new kinds of questions, more powerful questions in many ways, more precise questions, maybe more global questions. The questions never stop, they just change”
- The understanding of extinction was essential for the development of evolutionary theory: the concept of extinction was relatively new in biology and only accepted shortly before Darwin, because if species are perfect and never go extinct, it is hard to have a notion of evolution—so accepting that species are imperfect and can disappear was a necessary precondition for Darwin and others.
“if you don't have extinction and if species are perfect, then it's hard to have a notion of evolution. And so, the idea of extinction was itself part... The understanding of extinction was really essential for Darwin”
- We are entering a brave new world where evolution can be designed and manipulated: people grow organs in dishes via tissue engineering, resurrect ancient proteins from 400 million years ago using computation and evolutionary history to test their activity, and use cheap, effective CRISPR-Cas genome editing across many species to modify development—introducing teleology into a process that was previously goal-less.
“we're definitely in this brave new world where we can manipulate things as well. The big game changer has been, obviously, genome editing, the CRISPR-Cas gene editing, which applies to so many species. It's remarkably effective, but it's also remarkably cheap.”
- The theory that ontogeny (development) recapitulates phylogeny (evolutionary history)—promulgated by Ernst Haeckel after Darwin—is wrong as a law of nature; while there are individual cases where development tracks evolutionary history, it is not universally true. What is true is that early embryos of very different species share many similarities because they begin from common stages of development.
“It turns out to be wrong though there are cases where it does that... it's certainly not true as a law of nature, like Haeckel and his contemporaries proposed.”
- Despite moving far beyond Darwin—incorporating genetics, DNA, and exceptions like information being traded among species via viruses—modern evolutionary biology stays centered on Darwin: natural selection remains the major mechanism of evolution and common descent remains the major pattern, while the Darwinian revolution was fundamentally a profound shift in how we see the natural world and our relationship to it.
“Natural selection is still the major mechanism of evolution. Common descent is still the major pattern we see in the history of life, albeit with some exceptions where you have information being traded among species.”
- Auguste Duméril's Mexican salamanders demonstrated that a single subtle change in development—whether or not metamorphosis is triggered by thyroid hormone—can produce animals so different (one fully aquatic with external gills and a finned tail, the other fully terrestrial) that they appear to be entirely different genera, showing how a small developmental timing change can produce sweeping changes across the entire body.
“this difference is really a simple shift of development, and it showed how a simple shift of development can bring about changes across the entire animal”
Biology is messier than physics; human bodies carry evolutionary trade-offs and mismatches with modern environments causing disease.
- Trade-offs are a huge part of human evolution: as a highly optimized, extreme organism, humans pay costs for our advantages—our flexible throat and neural circuits enabling speech set us up for dangerous sleep apnea found in no other animal, our large brains consume enormous energy, walking on two legs causes problems, and our cognitive capacities for language may have been traded against other cognitive abilities like the short-term memory at which chimpanzees excel.
“our ability to talk comes at a huge trade-off... we are the only animals that suffer a particular kind of dangerous sleep apnea, and the reason for that is we have a very flexible back of our throat and a set of neural circuits that play a huge role in our ability to make sounds”
- Biology is a world of exceptions and evolution is characterized by a beautiful messiness—the mess is the message—as opposed to the cleaner frameworks of physics.
“It's a beautiful messiness. The mess is the message”
- Many human bodies are disconnected from the environments we now inhabit: we evolved from highly active ancestors but live sedentary lives in built environments, so our metabolism is ill-equipped, and leading causes of death—cardiovascular disease and cancers—are partly products of living in our modern world and living past 50.
“We evolved from highly active ancestors, and most of us are relatively sedentary... when you think about the leading causes of death, cardiovascular disease, cancers, and so forth, these are all parts of the trade-off of being human and living in our modern world”
Ancient viruses became integrated into the genome and repurposed for essential functions like memory, demonstrating evolution's creative reuse.
- About 8% of the human genome consists of defunct viral sequences—a graveyard of ancient viruses that attacked and became part of our genome before being knocked out—while only 2% of the genome is protein-coding genes.
“in humans, about almost 10%, about 8% of our genome, of our entire genome are defunct viruses. Viral sequences that no longer function, but they're there. It's like a fossil graveyard”
- The Arc gene, essential for memory in mice and people (its mutation associated with dementia and schizophrenia), produces protein spherules that look like the HIV capsid; Jason Shepherd showed Arc is actually an ancient virus that invaded a distant ancestor and, instead of being knocked out, was domesticated and repurposed—the capsule that protects viral genetic information cell-to-cell now carries genetic material neuron-to-neuron to make memories.
“what Arc is, is it's an ancient virus that invaded a distant ancestor of people. And that virus, instead of being knocked out by the genome, it was domesticated, it was repurposed.”
Traits once thought distinctly human have deep evolutionary antecedents in other animals, and evolution is a branching bush, not a linear ladder.
- Hiccups are a spasm of the phrenic nerve representing a stereotypical neural pattern inherited from our evolutionary past; one theory holds the same pattern occurs in frogs and tadpoles, which use a hiccup-like neural response to breathe with water, illustrating that traits we think are distinctively human have antecedents in other animals.
“That pattern occurs naturally elsewhere in the wild. In one place where we see it is in frogs and tadpoles, which use a form of a hiccup and that neural response to actually breathe with water.”
- The concept of 'missing links' is partly a misconception: while links among species do exist and are found, the underlying narrative wrongly assumes evolution is a linear ladder of progress, when in reality evolution is a hugely branching bush with many twists, turns, and reversals going in many directions.
“the path of evolution is often very unpredictable, it has lots of twists and turns and goes backwards and forwards and it's not a straight path, a linear ladder, it's a hugely branching bush of change that goes in many different directions”