Category: genetics

Colossal and Trump administration cooperate to sequence and store DNA of endangered species

June 28, 2026 • 10:15 am

Carl Zimmer and Catrin Einhorn are the authors of a new article in the NYT about our old friend Colossal Biosciences, which you’ll remember as the outfit in Texas that has promised to “de-extinct” the woolly mammoth, the thylacine (marsupial “wolf”), the dodo and the moa, after having claimed that they’ve already de-extincted the “dire wolf”.

As I’ve written at length here (and in an article in the Boston Globe), Colossal has not de-extincted anything. It simply edited 14 genes in a gray wolf cell, and then put that cell into the nucleus of a domestic dog egg. What came out were three slightly tweaked gray wolves, white in color and, Colossal says, larger than  normal wolves. But 14 changed genes in a genome of about 20,000 protein-coding genes, and having 2.5 billion DNA bases, does not turn a gray wolf into a dire wolf. Their response was that a dire wolf is anything that you think resembles a dire wolf, no matter how much. That is disingenuous.

I lost respect from Colossal when they decided to double down on their claim that they’ve brought something back from extinction, which they surely have not.  And their claims that they will release these things into the wild—their ultimate aim—is ridiculous. The three faux white dire wolves (I doubt the original was even white) are kept secretly on an enclosure somewhere in the West, with only a few toadying journalists or donors allowed to visit them.

Likewise, Colossal’s promise to give us woolly mammoths by 2028 is unbelievable, for they won’t be able to put an engineered Asian elephant egg into the endangered Asian elephant, much less produce a creature that has more than a minute fraction of mammoth DNA. On top of that, Colossal says their aim is to release these faux mammoths on the tundra, which won’t happen, and that when they do so, it will help with global warming since the furry elephants’ trampling on the permafrost will prevent release of carbon dioxide into the atmosphere, ameliorating global warming.  Gullible donors like Paris Hilton, Tiger Woods, and Tom Brady have swelled Colossal’s coffers by $400 million, and it’s now worth, notes the article below, more than $10 billion.

I think that when Colossal realized it couldn’t make good on its de-extinction promises, it started investing in other projects.  One of them is described in this article in the NYT (click below or find it archived here). What they propose to do, with the promised help of the Trump administration, is save the DNA from endangered species.  Now this project has its good aspects, for if Colossal sequences a lot of new genomes and publishes the sequences (which it promises to make public), we could learn quite a bit about evolution. And the American taxpayer doesn’t have to foot the bill for any of it.  But Colossal has no experience in “biodiversity banking” of this sort, even though nonprofit conservation organizations like the San Diego Zoo Wildlife Alliance has been doing it for over half a century (Colossal is decidely a for-profit company). The San Diego noprofit has in fact created clones of black-footed ferrets, a highly endangered species, from biobanked material, so at least it has something useful to show for its efforts.

Further, if Colossal is doing this for “de-extinction” purposes, and will retain sole possession of the material, as it will do, then it is preventing other organizations or scientists from using what is “banked.” The U.S. government has no business partnering with such an enterprise.  I don’t worry about de-extinction because that is (pardon the pun) a dead issue. But the concentration on biobanking may, as the authors note, “erode support for on-the-ground conservation,” which mainly involves saving existing habitat and keeping humans from destroying new habitat.

A few quotes from the article, which, as science journalism should, maintains a neutral viewpoint while emphasizing both pros and cons:

The Trump administration and a company that is promising to bring long-gone animals back from extinction announced a partnership on Thursday to preserve cells, tissue and DNA from threatened and endangered species.

The company, Colossal Biosciences, said its goal was to store samples from every animal and plant protected under the Endangered Species Act, which includes more than 2,300 listings worldwide.

As more species face the risk of extinction, scientists see such biobanks as a critical backup. But concerns are also growing that the rise of genetic engineering and efforts to revive extinct species will erode support for on-the-ground conservation, which often requires protecting habitat from drilling, mining and other development.

The announcement comes as the Trump administration has been rolling back protections on land and water, including through actions to weaken the Endangered Species Act, in favor of expanded oil and gas exploration, commercial fishing and other economic activities.

“This partnership brings together the scientific expertise of the U.S. Fish and Wildlife Service and the ingenuity of the private sector to develop new tools that can help recover species, preserve critical genetic resources, and strengthen the future of wildlife conservation,” Doug Burgum, the interior secretary, said in a statement.

Under a memorandum of understanding, Colossal and the Fish and Wildlife Service will collaborate to identify high-priority actions, and the government will provide a list of which species it wants to prioritize.

Well, I’d prefer that a consortium of scientists decide which species should be prioritized, and I’d prefer that the material be given to the San Diego Zoo organization rather than to Colossal, which will have sole use of the material and is a for-profit organization. The agreement is supposed to run for five years, and that Colossal gets to keep all the samples it collected with its own funding, equipment, or personnel”, which means pretty much all the samples.

Colossal has been busy doing other stuff, too:

After beginning its de-extinction efforts, Colossal branched out into biobanking. In February, the company announced a partnership with the United Arab Emirates to build what it calls a BioVault in Dubai, intended to store cell and tissue samples from more than 10,000 species.

Why Dubai? Why not store all the material in one place? Who knows? And they clone pets!

[Colossal] currently gets revenue from cloning pets and horses through a company it acquired last year, and claims to have future sources of revenue from licensing technology it develops for its de-extinction projects.

The article notes some criticism of Colossal’s proposal, too (I’m not quoting the criticism of the “de-extinction” endeavors, which the article also mentions):

But some conservation biologists expressed worries about depending so much for the long-term guardianship of precious samples on a private company.

“It seems like a bit of a risk for the U.S. government to place biomaterials in a for-profit company that doesn’t have a very long track record,” said Oliver Ryder, a conservation geneticist at the San Diego Zoo Wildlife Alliance, which operates a storage effort called the Frozen Zoo that has been preserving cells for about 50 years.

and

Gabriela Mastromonaco, chief science officer at the Toronto Zoo, called the U.S. plan laid out in Thursday’s announcement hugely ambitious.

“To collect every threatened and endangered species, that is a massive endeavor,” she said. “That means tracking, trapping, immobilizing, and getting your hands on a lot of animals.”

She expressed concern that the initial announcement was short on planning details that would be standard in many other nations.

. . . Dr. Mastromonaco of the Toronto Zoo said the announcement left many questions unanswered, such as how Indigenous communities would participate in decisions about the program and the rules for who gets to use the samples for research. She said she was addressing these questions herself as Canada develops its own plan for biobanking wild species.

and

Concerns that genetic engineering would replace critical conservation work heightened when Mr. Burgum, the interior secretary, celebrated the company’s announcement on X, writing that “the marvel of ‘de-extinction’ technology can help forge a future where populations are never at risk.” The Fish and Wildlife Service is part of the Interior Department.

Colossal executives emphasize that their efforts are intended to add to conservation strategies, not supplant the important work of protecting habitat.

I guess Colossal needs government cooperation since that’s required to collect DNA samples from endangered species. But if Colossal is dong this “for public good and impact” as Colossal CEO Ben Lamm has said, why do they retain the sole right to use the material? Even if it’s collected by Colossal, the permission to do so has to come from the U.S. government, and we should not be entangled with a private, for-profit company that will store material only it can use.

I see Colossal as having provided some valuable knowledge, but also largely as a pack of grifters, making promises they cannot keep and distorting what they have done.  In my view they should stick to cloning Fido and Fluffy for rich pet-owners who want to “de-extinct” their postmortem pets.

h/t: Don

Readers’ wildlife photos

June 3, 2026 • 8:30 am

Leucism, the absence of pigment in all or parts of the body in animals, is a genetic condition often mistaken for albinism (leucistic animals havenormally pigmented eyes).  It’s found in all sorts of animals, from reptiles to mammals, and Scott Ritchie has spotted it in Australian ducks.  Scott sent some pictures, which you can enlarge by clicking on them, and his captions are indented:

The leucistic Plumed Whistling Duck (Dendrocygna eytoni), is back at Hasties Swamp, Queensland, the white one in in middle.  We  have seen it for at least 2 years running. And “he/she” appears to have busy, with at least one (several white light feathers head and breast), and perhaps 3 (2 based on “forehead” feathers) individuals showing leucicism traits. It’s interesting that they were hanging together at the log to the left of the hide.

This last picture is of normal-type duck:

 

More on Colossal’s futile efforts to “de-extinct” ancient giant birds

May 20, 2026 • 10:15 am

After announcing that it would “de-extinct’ the Woolly Mammoth, and that it had in fact “de-extincted” the Dire Wolf, the company Colossal Biosciences is now making big noises about its effort to bring back extinct big birds: the giant moas of New Zealand (driven extinct by humans around 1300 AD) and the dodo of Mauritius (also killed off by humans in 1662). These are big birds (dodos weighed from 20-40 pounds, moas, of which there were nine species, from 55 to 600 pounds), and this fact alone makes it hard to de-extinct them.

But, as with the “woolly mammoth”, the concept of bringing back extinct species is impossible given our current technology, and saying that you can is grossly misleading.  And that’s for several reasons (the indented bits below are mine):

1.) You need the DNA of the extinct species if you’re going to create a simulacrum of it by injecting bits of the extinct animal’s DNA into the genome of a modern relative.

2.) You need to know what the DNA segments you have actually do in the animal, and how sequences differing between it and the donor genome can produce an animal with some traits of the original species. Where are the “big genes” in a mammoth, for example?

3.) Species are not “de-extincted”: what happens is that a living relative is genetically engineered by putting in a few bits of ancient DNA to create a “partial hybrid” that superficially resembles the ancient species. This involves finding and inserting a few bits of ancient DNA that you think will make the donor species look more like the extinct one. For example, I think fewer than 20 genes were engineered into a gray wolf genome to make the “dire wolf”. These included both dire wolf genes and mutant genes of modern dogs inserted into a gray wolf genome. The tweaked embryo was then implanted in a domestic dog. It’s important that you (and the press) realize that the ancient species is not brought back; what we get is a modern species that looks a bit like the ancient species. (See my post on the “dire-ish wolf” here.)

4.) There are problems with rearing the “tweaked” (I won’t call it “de-extincted”) species.  We cannot artificially inseminate elephants with elephant genomes that have been engineered for hairiness and bigger tusk. We don’t have the ability to do this (though we might in the future), the embryos might not develop properly, and the mother is unlikely to take care of them. This is why Colossal has spoken of using “artificial uteruses” to rear the tweaked “mammothy” elephant.

5.) For giant birds like the dodo and moa, you need to be able to rear the tweaked species—presumably adjusted to be larger than its surrogate relative—in big eggs. Because those eggs don’t exist, they have to be made somehow. This week Colossal announced the creation of 3-D printed eggshells that could be used to contain a chicken embryo that develops to term.  But of course hatching is one thing, and rearing is another. What mother will rear a tweaked “dodolike” bird.  The closest relative of the dodo is the Nicobar pigeon, a bird considerably smaller than the dodo (the pigeon weighs about a pound). The closest living relative of the moas is the tinamou, which weighs about as much as a big chicken: five pounds max.  Rearing such birds to maturity is a serious problem, even if they were full dodos or moas rather than tweaked pigeons or tinamous.

6.) Colossal has announced that a crucial part of “de-extinction” is “rewilding”: releasing the tweaked animals back into nature to restore their niche.  This is one of the most questionable parts of the whole enterprise.  The tweaked hairy elephant, for example, should be released on the tundra (and in groups of individuals, which is yet another problem, as you need to engineer more than one hybrid). That tundra doesn’t exist in the form it did in the past, and, of course, the tweaked hairy elephant has to have all genes necessary to seek out and use the food that a real woolly mammoth would eat, as well as genes for preferring as a mate others of its kind. It has to be able to survive extreme cold. We don’t know what genes these are! All we have are DNA sequences.

An example of the problems is Colossal’s announcement that it had “de-extincted” the Dire Wolf. It hadn’t: it engineered a gray wolf with about 20 inserted genes taken from both wolves and domestic dogs, producing a whitish wolf that seems a bit larger than gray wolves.  Three of these creatures were made. Not only were they not released in the wild, but they are sequestered in a secret and tightly-controlled fenced area that is off limits to all but selected journalists.

All the brouhaha, then, is misleading. We don’t get extinct species back, we may not even get “tweaked” species back, and they are very unlikely to ever see the wild again. I discussed many of these problem in an op-ed last year in The Boston Globe (archived here). See also the New Scientist article below.

Because Colossal has misled the public—they originally said they’d de-extincted the dire wolf, then retracted that claim, then reinstated it, saying that if it looks like a dire wolf, it is a dire wolf—each time they accomplish something they tout it as a huge advance towards real de-extinction. After all, they have to keep their rich investors and the public happy.

The latest Colossal announcement, which came through the mail, is that of their developing an artificial chicken eggshell. The problem is (see below the fold) that this has already been done by others some time ago. A further problem, of course, is that this is only a minor issue in the problem of putting dodo-like or moa-like embryo in an artificial egg. Here’s Colossal’s announcement, and note the emphasis on “de-extinction”:

BREAKTHROUGH: De-Extinction Just Got Its Egg
Step inside the beginnings of life as Colossal Biosciences hatches live chicks from our new artificial egg.
This huge advancement is foundational to our de-extinction of the South Island giant moa, whose eggs were around 80x the volume of a chicken’s. No living bird could possibly hatch one. So we built an artificial egg that will.
Watch a real chick embryo develop inside the artificial egg. Get a full breakdown of every feature. And see how this breakthrough is opening new doors for avian biotech research and bird conservation.
You’ll want to see the ending.
Meet the Colossal artificial egg. Nature spent millions of years perfecting the original. We just made our own, and hatched some beautiful and healthy chicks.
Here’s how it works:
🥚 Egg-shaped frame: a lattice shell that gives the whole system its structure and protection.
🌬️ Colossal membrane: the secret weapon. A bioengineered, gas-permeable layer that matches a real shell’s oxygen transfer, so O₂ flows in and CO₂ flows out exactly the way nature does it.
👁️ See-through build: the largely transparent design that lets us watch development in real time. This is critical for research and for de-extinction, where visually confirming milestones and the gene-edited traits we’ve put back is everything.
📏 Modular scale: the platform will stretch to fit eggs of any size, including the South Island giant moa egg, roughly 80x the volume of a chicken egg.
Extinction doesn’t have to be the end. And this is just the beginning.
Avian de-extinction is getting wild.

Here’s a breathy, chest-thumping video, accompanied by triumphant music, making it seem that the problem of de-extinction is on the way to being licked:

This is an achievement, of course, but to me it’s not a substantial step towards getting back moas and dodos. as it’s not that new.

And of course the press has picked it up, but this time they are careful to quote Colossal’s many critics as well as its chief propagandist, Ben Lamm. The Times of London talks about the eggshell as a step in resurrecting moas, using emus as surrogate moms. Click headline below to read:

Excerpts from The Times piece:

Colossal Biosciences, a Texan biotechnology firm, has developed a shell-less system it says is capable of supporting a bird embryo from early development through to the point of hatching.

So far the device has been used to produce baby chickens. The end goal, the company says, is to deploy a much larger version to resurrect the moa, whose eggs were about 80 times the volume of a farmyard hen’s.

. . . . So far the device has been used to produce baby chickens. The end goal, the company says, is to deploy a much larger version to resurrect the moa, whose eggs were about 80 times the volume of a farmyard hen’s.

. . .Eventually, the hope is that emu cells can be edited, introducing genetic changes that would make any resulting animal more moa-like. The hurdle then would be where to grow an embryo. According to Colossal, the eggs of the South Island giant moa were roughly eight times the volume of an emu’s. No living bird would be large enough to play mother to it.

This is where the artificial egg would come in. Colossal says the device could be scaled up, allowing embryos of much larger birds to develop in a controlled chamber. It claims this could remove the need for a living surrogate mother and make it possible to incubate embryos at sizes no modern bird can manage.

But they quote critics!

. . .Critics say such claims need careful handling. To recreate a mammoth, for instance, Colossal plans to alter the genetic code of an Asian elephant.

Even if that succeeds, sceptics argue the result would not truly be a mammoth, but an elephant engineered to have some mammoth-like traits, such as shaggy hair and extra fat reserves.

The same issues apply to the moa. The project, which is being backed by Sir Peter Jackson, the film director behind the Lord of the Rings trilogy, plans to compare ancient DNA from the extinct species with living relatives such as emus and tinamous to work out which genetic features helped make a moa a moa.

Well, there’s the rub! But Ben Lamm is always around to give the necessary donation-promoting optimism:

Ben Lamm, chief executive of Colossal, said: “Restoring species like the South Island giant moa isn’t just about reconstructing ancient genomes and editing [primordial germ cells, which eventually become sperm or eggs] — it requires building an entirely new incubation system where no surrogate exists.”

He added: “It’s a major milestone for Colossal and a foundational technology for our de-extinction toolkit.”

Again, I’m not saying that the artificial egg is not of any value. I’m just saying that insuperable problems remain with bringing back moas (or dodos).

Here’s a tweet that Matthew sent me, which called my attention to a New Scientist article that, mirabile dictu, strongly criticizes the de-extinction program as a whole:

Colossal says its "artificial egg" will help it bring back the moa, which had larger eggs than any living birds 🧪But it's really just an artificial eggshell, and even clingfilm will work – sort of – as an artificial eggshell. Plus there's the yolk problem…www.newscientist.com/article/2527…

— Michael Le Page (@mjflepage.bsky.social) 2026-05-19T12:19:10.242Z

The article at New Scientist can be found by clicking on the screenshot below, or finding it archived here:

A few Q&As from the piece:

Is this the first-ever artificial bird egg?

Colossal does use the term “artificial egg” in its press release, but it is really just an artificial eggshell. Either way, it isn’t a first – in fact, it’s possible to remove chicken eggs from their shells and hatch them from anything from plastic cups to cling film. However, the survival rate is usually low because, without an eggshell, the developing chicks may not get enough oxygen. A number of teams around the world have been working on more sophisticated so-called ex-ovo approaches.

How much better is it than cling film?

Colossal claims its silicone membrane is better than existing ex-ovo methods because it allows oxygen through at the same rate as a chicken eggshell and doesn’t require additional oxygen. However, it hasn’t released any experimental results to back this up. “I would love to see what the numbers are on efficiency,” says Ben Novak of non-profit wildlife conservation group Revive & Restore. “How many of these chicks hatch versus how many don’t?”

Colossal doesn’t publish much of the data that would enable scientists to see exactly what it did, which genes it used, and what the results are. Three more issues:

Does this mean we could create a giant artificial moa egg?

Even if Colossal’s approach does work well for chicken eggs, it won’t necessarily work for larger eggs. Larger eggs might need shells with different properties because of their lower surface-area-to-volume ratio, but this could probably be solved by tweaking the permeability of the membrane. Making a big egg also requires more than just a big eggshell. Moa eggs were up to 24 centimetres long and 18 cm wide, so they contained a lot more egg white and yolk than the eggs of living birds. Adding more egg white should be relatively straightforward. Chickens have been successfully hatched in the egg white from turkeys, says Novak, which suggests it won’t matter much what animal’s egg white is used.

How much better is it than cling film?

Colossal claims its silicone membrane is better than existing ex-ovo methods because it allows oxygen through at the same rate as a chicken eggshell and doesn’t require additional oxygen. However, it hasn’t released any experimental results to back this up. “I would love to see what the numbers are on efficiency,” says Ben Novak of non-profit wildlife conservation group Revive & Restore. “How many of these chicks hatch versus how many don’t?”

What about the extra yolk required?

That’s more of a problem. Each egg yolk is a single cell, meaning ostrich yolks are the largest single cells found on the planet. To make a yolk larger would require penetrating the cell membrane and injecting more yolk, says Novak, which would likely cause the cell to burst. So it might be necessary to somehow enlarge the cell membrane so it can hold the extra yolk. Which is probably doable, given enough time, effort and money.

And the $64,000 question (and answer):

If all the problems with making really big eggs are solved, could we then bring back the moa?

Absolutely not. The big problem is that DNA breaks up into smaller and smaller pieces over time, and all nine species of moa went extinct around 600 years ago. It’s never going to be possible to get a complete, working copy of a moa genome and produce a living animal genetically identical to any of the extinct species. Even the human genome wasn’t completed until 2023.

. . . . Is this what Colossal aims to do with the moa?  [i.e., tweaking existing species with a few gene edits]

Colossal hasn’t revealed its plans, but based on what it did with grey wolves and “woolly mice”, its aim will likely be to tweak a few genes in the emu to create something a bit like the moa in appearance, even if that means making genetic changes not present in the moa genome. Five of the 20 gene edits made to the grey wolves were changes not found in the dire wolf genome, for instance. Nic Rawlence at the University of Otago in New Zealand doesn’t think Colossal is close to achieving even this for the moa. “The development of a genetically engineered emu and calling it a moa for no good conservation or ecological reason, but rather an ecotourism venture, is still a long way off,” says Rawlence. Nor is the idea of bringing back the moa universally welcomed, he says. “There is widespread Māori and public opposition in Aotearoa New Zealand.”

Grok agrees with all this when I asked it the question below:

From Grok:

Question I asked it: Will Colossal be able to bring the moa and the dodo, that is, “de-extinct” them?
Grok’s answer:
No, not in the sense of resurrecting the exact extinct species—but Colossal Biosciences could plausibly create moa-like and dodo-like engineered proxy birds (genetically modified versions of living relatives) within the next 5–15 years, assuming continued progress. True de-extinction (reviving genetically identical copies of the original species) is not possible with current or foreseeable technology.

There’s more at New Scientist, but you can read it at the link or the archived link. I hope I’ve shown that Colossal is engaged in misleading the public (and I can’t help but think it knows this, since it de-emphasizes the “tweaking” part), and that you’ve learned some of the problems with its “de-extinction” claims.

Below the fold I’ve put Grok’s answer to my question about whether previous workers had reared chicken eggs using artificial “shells” previously. The answer is “yes,” though Colossal’s expensive shell is more sophisticated. I cannot vouch for the accuracy of Grok’s answers, but of course it tells you how to investigate them.

h/t Pyers.

Click below to read more:

Continue reading “More on Colossal’s futile efforts to “de-extinct” ancient giant birds” →

A short obituary of J. D. Watson in PNAS

February 18, 2026 • 9:45 am

The Proceedings of the National Academy of Sciences finally published an obituary of J. D. Watson, who died in November of last year. (Nathanial Comfort has written a biography of Watson that will be a good complement to Matthew’s biography of Crick; Comfort’s book will be out at the end of this year or the beginning of 2027.)  You can access the PNAS obituary for free by clicking on the screenshot below, which is a good summary of Watson’s accomplishments (and missteps) if you don’t want a book-length treatment.

Most laypeople, if they know Watson’s name, probably know just two things. First, he and Crick co-discovered the structure of DNA, one of the great findings of biology. Second, Watson was demonized, and fired as director of the Cold Spring Harbor Laboratories, for making racist comments.  Both are true. Yes, Watson was a racist, as I discovered from talking to him for an hour and a half (see below), but he was also a brilliant scientist who did far more than just the DNA-structure stuff. The article describes some of his other accomplishments and I quote:

DNA was not the only structure that Watson solved at Cambridge. Using X-ray crystallography, Watson determined that the coat protein subunits of Tobacco Mosaic virus (TMV) were arranged helically around the viral RNA, although he could not detect the RNA (5). Two years later, Rosalind Franklin, now at Birkbeck College with J. D. Bernal, published the definitive study on the structure of TMV (6).

Watson left Cambridge in 1953 to take up a fellowship with Delbrück at the California Institute of Technology. He joined forces with Alex Rich in Pauling’s laboratory to work on the structure of RNA, but RNA gave fuzzy X-ray diffraction patterns and provided no clues as to what an RNA molecule might look like. Watson was not happy in Pasadena and, with the help of Paul Doty, was appointed an assistant professor in the Department of Biology at Harvard. However, he first spent a year in Cambridge, United Kingdom, before moving to Cambridge, Massachusetts.

Watson and Crick teamed up again to study the structure of small viruses and proposed that as a general principle, the outer protein coat of these viruses was built up of identical subunits. Franklin was also studying small viruses, and she and Watson exchanged letters, and she asked Watson and Crick to review drafts of her manuscripts.

At Harvard, Watson, his colleagues, and students made many important findings on ribosomes and protein synthesis, including demonstrating, concurrently with the team of Sydney Brenner, Francois Jacob, and Matt Meselson, the existence of messenger RNA. Watson’s contributions are not reflected in many of the publications from his Harvard laboratory. He did not add his name to papers unless he had made substantial contributions to them, thus ensuring that the credit went to those who had done the work. These papers included the discovery of the bacterial transcription protein, sigma factor, by Watson’s then graduate student Richard Burgess, along with Harvard Junior Fellow Richard Losick. At Harvard, Watson also promoted the careers of women, notably providing support for Nancy Hopkins, Joan Steitz, and Susan Gerbi. He also contributed to the split in the Department of Zoology due to his contempt for those working in the Department who were antireductionists.

 

In his last scientific paper (7), published in 1972, Watson returned to DNA. In considering the replication of linear DNA of T7 phage, he pointed out that the very ends of a linear DNA molecule cannot be replicated, the “end replication problem” which is solved in eukaryotes by telomeres. (Watson’s work was predated by Alexey Olovnikov who had published the same observation in 1971 in a Russian journal.)

Note the contributions Watson made, along with collaborators, at Harvard, and note as well that he did not put his name on publications unless he made “substantial contributions to them.”  I did that, too, and I inherited that practice from my Ph.D. advisor Dick Lewontin, who inherited it from his Ph.D. advisor Theodosius Dobzhansky, who inherited it from his research supervisor at Columbia and Cal Tech, the Nobel Laureate T. H. Morgan.  This is a good practice, and I never suffered from keeping my name off papers, for the granting agencies care only about which and how many papers come from an investigator’s funded lab, not how many his or her name is on.  I’ll digress here to say that this practice has almost died out, as people now slap their name on paper for paltry reasons, like they contributed organisms or other material.  The reason is the fierce competition for funding and credit.

Watson went on to write influential textbooks, trade books (notably The Double Helix) and headed up the Human Genome Project, from which he ultimately resigned. Finally, he ran the Cold Spring Harbor Laboratory, which he did very well until the racism scandal broke, rendering him ineffective.

Witkowski and Stillman don’t neglect the dark side of Watson:

In the late 1990s, Watson gave seminars, notably at the University of California Berkeley, where he expanded on research on the hormone POMC and related peptides and made inappropriate and incorrect observations about women. In October 2007, he made racist remarks about the intelligence of people of African descent, and, damagingly for his fellow employees at CSHL, stated that while he hoped that everyone was equal, “people who have to deal with black employees find this not true.” The CSHL Board of Trustees dissociated the institute from Watson’s comments, and he was forced to step down from his administrative position as Chancellor. The matter resurfaced in January 2019 when Watson was asked if his views on race and intelligence had changed. His answer was unequivocal: “No, not at all.” The Laboratory’s response was immediate, relieving him of all his emeritus titles. Watson and his family, however, continued to live on the CSHL campus.

They conclude this way:

Jim’s remarkable contributions to science and society will long endure—for the scientists using the human genome sequence, for students using Molecular Biology of the Gene and for readers of The Double Helix, and for reviving Cold Spring Harbor Laboratory. He was a most amazing man.

Here’s a photo of Watson and me when he visited Chicago in 2013 to introduce the Watson Lectures that he endowed for our department. Do read the cool story about how those lectures came about in my post “Encounters with J. D. Watson“.

Michael Shermer interviews Matthew Cobb on his Crick biography

January 18, 2026 • 9:45 am

Here we have an 83-minute interview of Matthew Crick by Michael Shermer; the topic is Francis Crick as described in Matthew’s new book Crick: A Mind in Motion. Talking to a friend last night, I realized that the two best biographies of scientists I’ve read are Matthew’s book and Janet Browne’s magisterial two-volume biography of Darwin (the two-book set is a must-read, and I recommend both, though Princeton will issue in June a one-volume condensation).

At any rate, if you want to get an 83-minute summary of Matthew’s book, or see if you want to read the book, as you should, have a listen to Matthew’s exposition at the link below.  I have recommended his and Browne’s books because they’re not only comprehensive, but eminently readable, and you can get a sense of Matthew’s eloquence by his off-the-cuff discussion with Shermer.

Click below to listen.

I’ve put the cover below because Shermer mentions it at the outset of the discussion:

On the origin of venom by means of natural selection

December 11, 2025 • 10:20 am

Many animals are venomous, but in most cases the exact proteins involved in causing pain or death are unknown, and even in those cases the genes producing them have not been identified, counted or mapped.  If you’re interested in the evolution of venom, what its precursors are, and how venomous animals avoid poisoning themselves, you have to know this kind of stuff.

A new paper in Proc. Nat. Acad. Sciences (click screenshot below to read for free, or find the pdf here) answered several of these questions in the venomous caterpillar of the mottled cup moth (Doratifera vulnerans), shown below.  It’s from Australia, and is described in Wikipedia this way:

It is known for its caterpillar having unique stinging spines or hairs that contain toxins, for which the scientific name is given that means “bearer of gifts of wounds”. Chemical and genetic analysis in 2021 show that its caterpillar contains 151 toxins, some of which have medicinal properties

That earlier paper, from 2021 and including some of the same authors as the one we discuss today, did indeed identify 151 proteins (peptide are bits of proteins or short chains of amino acids) that were in the toxins, but did not know which genes produced them, how the genes were arranged, what the closest relatives of the genes were, and how many of the 151 “toxins” were really toxic (the word “toxin” there and in the present paper do not mean that the substances were toxic, but that they were simply a component of the extracted toxins). However, the authors, some on the paper I’m highlighting today, did identify two genuine toxins that caused pain: the peptides Dv12 and Dv11.

Look at this thing! It’s clearly aposematic, meaning that it has bright warning coloration that predators can recognize and learn to avoid. And you can see those nasty spines.  In the earlier paper they extracted toxins from related species and tested them by injecting them into mice tails, guinea pigs, and human volunteers. That earlier paper also adds this about the species name:

This species, whose binomial name etymologically means “bearer of painful gifts,” is a common culprit of caterpillar envenomations in Australia.

That means that many Aussies get stung by these things, probably inadvertently. Would you touch an animal that looks like this?:

Photo by Fir0002,  Creative Commons Attribution-Share Alike 3.0 Unported license.

On to the new paper, and I’ll try to be brief as it’s long and complicated.

1.) First, the authors sequenced the entire caterpillar genome (remember, it’s the same as the adult moth genome).

2.) Then, knowing the sequences of the proteins known from previous work on toxins, they could find the genes producing them by matching the protein sequence to the DNA sequence that could produce these proteins. Of the 151 proteins in caterpillar venom known from the prvious work, they mapped 149 of them to 115 sites in the genome

3.) Of the 115 sites, 35 were products of single genes, while 80 (70%) of the total, were members of gene families consisting of two or more similar genes (sometimes many genes) with similar sequences.  Here’s a map of the “toxin gene” locations on the insect’s 13 chromosomes. The blue dots are the genes existing in single copies, orange dots are clusters of genes previously grouped together by protein-sequence similarity, and pink dots are genes that were newly identified, surely as part of gene families, in the present study. This conclusion comes from their sequence similarity and they physical grouping on two chromosomes.(The size of the dots indicates the number of genes that are part of a contiguous group. Click to enlarge:

So we know that genes found in venom are very often the product of gene duplications, either of single genes becoming two (this can happen via unequal crossing-over during meiosis or by other methods), producing two initially identical genes side by side or whole groups of them (“tandem duplications”). Once a gene has been duplicated, the original copy can then keep its original function, while the other copies, not being “needed,” are free to evolve other functions. Many genes we’re familiar with, like our own globins and immunoglobulins, evolved by gene duplication followed by divergence of the duplicated copies.

Where did the genes making venom proteins come from? This is the key evolutionary question answered here and, to some extent, in the previous paper. They evolved from ancestral genes in the moth’s immune system that evolved to attack microbes, the so-called “antimicrobial peptides” (AMPs), also known as cecropins. The ancestral AMP proteins, nearly identical to their original form and function, kill bacteria (prokaryotes) by disrupting the bacterial membranes. Insects still need to kill microbes!

Clearly, the proteins in venom have evolved by natural selection modifying ancestral genes used to kill bacteria. Now they are used to repel predators. Natural selection causing this divergence was implicated by looking at sequence differences, as there are ways of showing what sequence differences evolve faster than expected under either the slower processes of genetic drift or “purifying” selection that conserves structure.  They found that most of the venom-adapted proteins that evolved from cecropins did evolve under natural selection, while the descendants of cecropins that retained their original anti-microbial proteins were under purifying selection to retain their sequence. It’s clear, then, that the insect still needs genes to attack bacteria. It’s just that some of them have been repurposed, often through gene duplication and divergence, to repel predators. (The authors have a way of assessing “pain” by measuring the increase in calcium concentration in cells grown in vitro and exposed to venom. This happens when the two investigated proteins are used.)

Here is a complicated family tree of cecropin genes in black used to kill microbes. The genes found in venom are in the red box (“venom adapted”). You can see that they are related to cecropin genes but branched off fairly recently (probably four or five million years ago). The venom genes are in the red box that I’ve added, and their relationship as being derived from ancestral AMP genes is very clear. (The “canonical” genes in green are antimicrobial proteins closely related in sequence to the venom genes.

So, now we know where the genes in venom come from. What we do not know is how many of those genes are essential in venom, either causing pain or doing other stuff that venom needs to do. At least two of them cause pain, but there are probably more, for they haven’t all been tested. And some of the other genes are probably involved in dismantling cell walls in potential predators. The authors tested several of the venom proteins and also found that, as in their AMP ancestors, they disrupt cell covering, in this case eukaryotic cell membranes.

Finally, the big question: If the caterpillar makes venom, why doesn’t it poison itself? Here’s how the authors answer that question (I’ve put the answer for this species is in bold).

Animals that produce toxins, either for innate immunity or as venom toxins, must employ strategies to protect themselves from toxicity. Such protective mechanisms include production of toxin inhibitors, storage in inactive form, mutations in their own ion channels that confer resistance, alteration of lipid bilayer compositions, and compartmentalization of toxins separate from body tissues. In the case of limacodid venom peptides, the venom is compartmentalized into the cuticle-lined venom reservoir inside venom spines, preventing the toxin from coming into contact with cells other than the secretory cells that produce them. Thus, compared to canonical cecropins, venom-adapted cecropins may also be released from pressure to avoid activity against animal cells.

There are other findings in the paper that will be of interest primarily to those studying genomic evolution. For example, many of the venom proteins still retain some weak antimicrobial activity, so the idea that genes completely lose their ancestral function when they gain a new one doesn’t hold in this case.

Below you can see the adult moth because, remember, they studied caterpillar venoms, and many of those genes are probably turned off in the adult. But adult and caterpillar carry the exact same genes, of course; their different bodies, physiology, and behavior rest on the differential turning on and off of these genes at different life stages. And that remains a big mystery: how do such different life stages evolve, with each step of the evolution being adaptive?

From The Australian Museum, photo credits at bottom (click to enlarge), image by Lyn Craggs.

 

Human and chimp genome comparison: apples and origins

December 7, 2025 • 10:00 am

How much genetic difference separates us from our closest relatives? The conventional wisdom about humans and our closest ape relatives (chimps and bonobos) is that we share 98% of our DNA. That’s a big similarity, and implies that if we lined up our genomes side by side, only about 2 out of 100 DNA bases would differ. This figure is often used to show that we have only a tiny genetic difference from our closest relatives. To quote W. S. Gilbert of Gilbert and Sullivan, “Darwinian man, though well-behaved, at best is only a monkey shaved.”  Well, the differences go farther than mere shaving.

The “98% similarity figure” is wrong. And it’s wrong for several reasons. First, most ape genomes (chimps, gorillas, orangs, etc.) have not been as thoroughly sequenced as was the human genome. A lot of the data that went into the 98% figure was missing.  Second, you can’t just compare genomes by lining them up and looking for differences in base pairs at similar sequences.

Why not? Because the notion of “similar sequences” is ambiguous and, sometimes, meaningless. Since we diverged from our ape ancestors, there have been a lot of changes in every species’ DNA that prohibit us from simply “lining up the genomes”.  Transposable elements have invaded some species but not others, bits of the DNA have been duplicated, so there are species that have sequences that are not homologous. Bits of the genome have been inverted (turned around and reinserted), causing big differences in sequence in previously similar sequences. Further, pieces of the DNA have been moved from one chromosome to another, so DNA sequences previously in the same place are now in another place, leading to a difference in total sequence.

All this leads to a substantially greater DNA divergence between humans and chimps than the 98% figure.  These extra genomic differences were sussed out by Yoo et al. in a Nature paper  from April of last year that you can read by clicking below (or find the pdf here).They did a much improved job in sequencing six of our ape relatives: the chimp (Pan troglodytes), bonobo (Pan paniscus), Western gorilla (Gorilla gorilla), Bornean orangutan (Pongo pygmaeus), Sumatran orangutan (Pongo abelii), and the siamang (Symphalangus syndactylus), an endangered species of gibbon from SE Asia.

First, the authors give a revised set of divergence times based on DNA differences between living species.  The human vs. chimp/bonobo species, for example, split from their common ancestor about 5.5-6.3 million years ago (mya), roughly in line with previous estimates. The divergence between humans and other African apes (gorillas) occurred between 10.6 and 10.9 mya, and that between humans and orangutans about 18.2-19.6 mya.

There is a ton of genomic information in the paper, including a lessening of the similarity between humans and chimps, but also specific information about what genes and regulatory bits of DNA differ among species. These differences suggest some some intriguing future research. I’ll mention just a couple, but will refer you instead to a long tweet below which shows why the human-chimp differences have increased. It’s an excellent tweet that you can read pretty quickly, though it doesn’t detail all the many differences that the researchers describe in the Nature paper, which is exhausting for those outside the field. There are also genes whose sequences changed very rapidly, suggesting that they were acted on by natural selection.

There are a gazillion sequence and structural differences revealed among the species, including 229 bits of ape DNA (all species) that have evolved rapidly and are thus candidates for natural selection. The paper also reveals parts of the DNA that have evolved especially rapidly in the human lineage since we split from chimps/bonobos. These regions are called HAQERS, and could be candidates for the Holy Grail of such work: seeing “what makes us human”. But that question is a bit misguided.

Nevertheless, the authors found one gene, ADCYAP1, that “is differentially regulated in speech circuits.” The implication is that the changes may have something to do with why humans are the only ape with syntactic spoken language, but that gene does a lot of other stuff, too, so I don’t take that implication seriously. The FOXP2 gene, which evolved rapidly in the modern human genome relative to other species, has mutations that impede people’s ability to speak, and I well remember when it was touted as “the language gene” that enabled humans to speak. But further research showed that the accelerated human evolution of the gene was an artifact, and that the normal function of the gene is manyfold, so nobody these days takes FOXP2 seriously as the “speech gene”. All claims should be regarded as caveat emptor.

There are also several genes that are not only unique to humans, but are “associated with human evolution of the frontal cortex”, suggesting these account for our big brains. The photo below comes from the tweet shown next, and its caption comes from that tweet. (The average chimp brain is about 400 g in mass—less than a third the mass of the human brain, which weighs in at 1300-1400 g in adults.)  Again, caveat emptor with regard to the two specified genes.

Figure 3. Radiograph illustrating cranial expansion in the human lineage, which is associated with increased neocortical growth – Chimpanzee skull (left), Modern Human skull (right).

Other genes that differ strongly among ape species involve those producing immunoglobulin, major histocompatibility products (MCH) and T-cell receptors, but especially immunoglobulin genes—involved in production of antibodies. Why have these evolved so rapidly within apes? Your guess is as good as mine, but suggests that reaction to antigens was an important element of ape evolution.

Here is the authors’ summary, and most of the paper will be of interest only to geneticists familiar with the argot (not necessarily me):

The complete sequencing of the ape genomes analysed in this study significantly refines previous analyses and provides a valuable resource for all future evolutionary comparisons. These include an improved and more nuanced understanding of species divergence, human-specific ancestral alleles, incomplete lineage sorting, gene annotation, repeat content, divergent regulatory DNA and complex genic regions as well as species-specific epigenetic differences involving methylation. These preliminary analyses revealed hundreds of new candidate genes and regions to account for phenotypic differences among the apes. For example, we observed an excess of HAQERS corresponding to bivalent promoters thought to contain gene-regulatory elements that exhibit precise spatiotemporal activity patterns in the context of development and environmental response99. Bivalent chromatin-state enrichments have not yet been observed in fast-evolving regions from other great apes, which may reflect limited cross-species transferability of epigenomic annotations from humans. The finding of a HAQER-enriched gene, ADCYAP1, that is differentially regulated in speech circuits and methylated in the layer 5 projection neurons that make the more specialized direct projections to brainstem motor neurons in humans shows the promise of T2T genomes to identify hard to sequence regions important for complex traits. Perhaps most notably, we provide an evolutionary framework for understanding the about 10–15% of highly divergent, previously inaccessible regions of ape genomes. In this regard, we highlight a few noteworthy findings.

The importance of the paper for now seems to be the presentation of the sequences and their differences rather than explaining the differences or their significance in ape adaptations—especially in humans—for studying adaptive hypotheses involves a lot of work for each single region that differs among species or evolved quickly. Nevertheless, useful questions have been raised—like why genes involved in the immune response changed so rapidly—that will be subject to future work.

I am not sure who runs the Origins Unveiled site dealing with evolutionary anthropology, but based on the clarity of the tweet below from that site (click on screenshot to see the tweet in situ), it deserves more followers. It’s only about a year old, which may explain the follower issue.

This tweet from September of this year explains why the 98% similarity between humans and chimps drops to 84.7% when you take translocations, inversion, duplications, insertions, and other genomic rearrangements into account. And these rearrangements are not necessarily trivial, for duplications can lead to divergent gene families, and insertions can act to regulate genes in a new way.

Again, click below and read; it’s short and lucid:

I’ve shown one figure from the tweet above: the brain differences. Below is another figure showing how the 99% similarity between humans and chimps has traditionally been calculated, requiring alignment of nearly identical but perhaps slightly different bits of DNA. All captions come from the tweet. This figure shows how they line up chimp and human sequences (you see the gross similarity), but also that here there’s been a single nucleotide substitution in one of the two lineages, rendering this sequence 92.3% similar. (This is a made-up sequence for purposes of illustration.)  When you did that with the whole genome comparison based on earlier data, you got about a 2% difference. The problem, as I said, is that we didn’t have great chimp (or any ape) sequences and there are parts that you simply couldn’t line up this way. And those parts, when compared among species, increase the genetic difference between us and our closest relatives.

Figure 1 — Simplified Mock Alignment Illustrating Nucleotide Sequence Similarity Between Chimpanzee and Human Genomes. Out of 13 positions, one substitution (single-nucleotide variant, circled in red) results in ~92.3% DNA similarity. This example demonstrates the methodology behind the misleading 98–99% human-chimpanzee DNA similarity figures.

Below is another figure showing how various rearrangements, insertions, deletions, and translocations reduce similarity, but I’ll show only four of the six parts of the figure, giving the captions for a-d. You can see how these changes make humans and chimps less genetically similar than previously thought (again, captions come from the tweet; click to enlarge).  These are also “mock alignments” meant for purposes of illustration, but they do show the kind of thing seen in the Yoo et al. paper:

Figure 2 — Simplified Mock Alignments Illustrating Structural Variation Between Chimpanzee and Human Genomes. Note: Structural variants are not taken into account when calculating the 98–99% Chimpanzee-Human DNA similarity figures.
( a) Insertions and deletions contributing to sequence divergence. Out of 34 positions, 3 indels (insertions circled in orange; deletions in yellow) result in ~91.2% DNA similarity. Note: These indels are relative, as without a suitable outgroup (i.e. gorilla), an insertion in one genome appears as a deletion in the other.
(b) Duplication contributing to sequence divergence. Out of 34 positions, a duplication of 12 bases (duplicated segment encircled in blue; original in purple) results in ~64.7% DNA similarity.
(c) Inversion contributing to sequence divergence. Out of 34 positions, an inversion of 11 bases (encircled in green) results in ~67.6% DNA similarity. Note: Although bases may match within the inverted region, they do not contribute to sequence similarity due to misalignment. Without a suitable outgroup (i.e. gorilla), it is unknown whether the inversion occurred on the chimpanzee or human genome.
(d) Translocation contributing to sequence divergence. Out of 34 positions, a translocation of 20 bases (encircled in brown) results in ~41.2% DNA similarity. Note: A translocation is a DNA segment that has been “copy and pasted” or “cut and pasted” from another part of the genome.

So, when you hear that we’re nearly genetically identical to our closest relatives, just say, “Wait a tick. Not all that identical.” We have about 15% difference in sequence, which is not trivial.

UPDATE: I’m aware now that creationists and IDers have been using this 85% to cast doubt on human evolution, our place in the ape family tree, and whether evolutionists are honest.  This is bogus: the 85% vs. 98% depends on two different methods of calculating similarity. Which ever method you choose (alignment vs. total genomic similarity), the same family tree of the great apes appears, with chimps/bonobos our closest ancestors, then gorillas a bit more distance, and then orangutans, and then other apes.  The point of this post is not to cast doubt on human or ape evolution, but to show different ways of calculating genetic similarity.