How the whale lost its genes

September 27, 2019 • 8:30 am

The evolution of whales, porpoises, and dolphins—the “cetaceans”—is well understood thanks to a plethora of fossils, mostly found in recent years (for a good general summary of the data, go here). Starting from a small, deerlike artiodactyl living around 48 million years ago (Indohyus may be related to the common ancestor of whales), this evolution proceeded rapidly, with the two groups of modern whales—Odontocetes (toothed whales, including dolphins) and Mysticetes (baleen whales), diverging only about 12 million years later. In other words, in a mere 12 million years—only about twice the time since we diverged from the lineage that led to modern chimps—evolution went from a terrestrial artiodactyl to a fully marine whale. That’s surely macroevolution, however you define it, and gives the lie to the creationist claim that major transitions aren’t seen to have occurred over time. (I discuss much of the fossil evidence in Why Evolution Is True).

Here’s a diagram of the evolutionary sequence of some of the forms, and the times they appear in the fossil record, taken from the UC Berkeley site Understanding Evolution. 

During this brief period, cetacean ancestors lost their hind legs and developed a fluke, the body became streamlined for swimming, body hair was lost (not needed in a fully marine whale), the nostrils evolved backward into a single blowhole, a lot of adaptations for diving evolved, as well as the ability of the species to collapse their lungs when diving, a layer of blubber evolved, and there were many other physiological and anatomical changes. These are described in a new paper in Science Advances (click on screenshot below, pdf here, reference at bottom of the post).

But the authors are not so concerned with the well-documented morphological changes, for they wanted to see what genes had changed, in particular, which genes in the ancestors of whales had been inactivated during whale evolution—inactivated because they were of no use to fully marine mammals.

To find these genes, the authors looked at whole genomes of cetaceans (bottlenose dolphins, killer whales, minke whales, and sperm whales) and compared them to the genomes of 62 placental mammals. They found 2472 genes in cetaceans that were broken: genes having deletions, stop codons, splice-site mutations, “frame-shift” mutations, and other changes that kept these genes from being expressed. They then carved out a subset of these genes that were not inactivated in 95% of the terrestrial mammals they used for comparison, so the broken genes were largely unique to all cetaceans. That took the sample of broken cetacean genes down to 350.

They then excluded genes already known to be broken in the cetacean lineage, including olfactory receptor genes (I discuss the loss of “smelling genes” in whales in WEIT) and keratin-associated genes, most involved in hair formation.

Finally, they excluded genes known to be broken in the closest living relative of whales—the hippopotamus, whose “broken genes” are not associated with a marine way of life. This left the authors with a final sample of 85 genes that were inactivated in all sampled whales (mysticetes and odontocetes) but not in their living relatives; these were presumably genes that got broken in the common ancestor of the two groups of whales, and whose broken state was passed on to all living cetaceans.

Why would a gene become inactivated in a group? Well, presumably because it’s not needed. But there are then two ways that a non-useful gene could become broken via the accumulation of inactivating mutations. First, an inactivation could be “neutral”: a gene that’s not needed and becomes nonfunctional may not have a selective advantage or disadvantage over the active form, and could eventually become “fixed” (present in all individuals in a population) via random genetic drift.

Alternatively, a broken gene could increase in frequency because it has a selective advantage over its functional competitors. That is, the non-production of a gene product could save energy that could be diverted to other functions, or it could reduce an unneeded organ or feature that could be damaged (both of these arguments have been used to explain why eyes largely disappear in cave animals who don’t “need” them). The authors posit that most of the broken genes in cetaceans accumulated by neutral processes, but it’s very hard to distinguish that scenario from an increase-by-selection argument, as this involves comparing DNA sequences and looking for a “signature of selection”: nearly impossible in such data.

But this is a side question. What’s important are two things. The first one I emphasized in WEIT:

1). The presence of nonfunctional genes in whale genomes—genes that are functional in their living relatives—is strong evidence for common ancestry of whales from terrestrial organisms and against any creationist scenario. It’s also evidence for macroevolution. You’d be hard pressed indeed to give reasons why an intelligent designer or a god would install useless genes in a genome that remain useful in the “outgroup” relatives. But this is exactly what is expected if whales evolved from terrestrial species in whose descendants those genes remain useful. But we’ve known about broken genes for a long time (e.g., olfactory genes), and IDers and creationists still can’t explain them.

2.) The broken genes give evidence about what the genes were used for in the ancestors, and why they weren’t needed in cetaceans. Thus, the authors looked at what the genes do when they’re functional, which helps tell us why they might not be needed in cetaceans. The broken genes fall into several classes; I’ll highlight just three:

a. Genes involved in blood coagulation.  When cetaceans dive, peripheral blood flow is reduced, making it more likely that damaging blood clots could form, especially when nitrogen microbubbles form in the blood (this is what causes “the bends” in divers). The authors found two genes involved in blood coagulation that were broken in whales. Like all of the broken genes, this scenario for why genes are inactivated is speculative, but can still prompt further research.

b. Genes involved in DNA repair. When tissues become short of oxygen, as when cetaceans are diving, but then get a surge of oxygen later, forms of “reactive” oxygen accumulate that can damage DNA. Cetaceans have lost an enzyme, POLM, that repairs DNA, but does so by inducing many errors in the repaired DNA. Since there are other less error-prone ways of repairing DNA, the authors speculate that the loss of POLM is a way to avoid a “mutagenic risk factor” in cetaceans, which are especially prone to damaged DNA.  The idea is that it’s better, if you’re susceptible to damaged DNA from diving, to get rid of a system that repairs quickly but makes errors, and rely instead on another system that repairs more slowly but with fewer errors.

c. Genes involved in melatonin biosynthesis. Whales, like ducks, sleep with only half of their brain at a time, with the other half active and awake to watch for danger and, in whales, to keep the animal swimming, surfacing, and breathing, maintaining body heat. The sides alternate over time so that the entire brain eventually gets a rest. (This shows how important sleep is, though we don’t yet know why.) Melatonin is a hormone synthesized by the pineal gland that helps keep animals entrained to daily (circadian) rhythms. The authors found that four genes involved in melatonin synthesis (AANAT, MTNR1A, MTNR1B, and ASMT) were inactive in cetaceans but not in their relatives.

The authors speculate that the loss of melatonin synthesis “helps decouple sleep-wake patterns from daytime,” as whales sleep with half their brains during both day and night. Further, since melatonin synthesis inhibits body-temperature regulation, its absence may help whales maintain their high body temperature in a chilly environment.

There were other pathways in which cetaceans showed broken genes, including those involved in transporting amino acids to the kidneys and genes expressed in the lungs, which may facilitate non-damaging lung collapse  that occurs in diving cetaceans. You can read about these in the paper; again, the reasons for their loss are plausible but speculative.

Finally, the authors looked at two other groups of aquatic mammals whose ancestors independently invaded the sea: manatees (sirenians, related to elephants and hyraxes) and pinnipeds like seals and sea lions (descended from terrestrial carnivores). Their goal was to see if there was independent “convergent” loss of similar genes between these groups and cetaceans. They found two genes, including AANAT, that were inactivated in manatees or pinnipeds, but not in their terrestrial relatives.

What does it all mean? As I said above, this paper gives further evidence for evolution in the form of dead genes, genes not needed in some groups of animals but needed (and “alive”) in their terrestrial relatives and presumably ancestors. This gives further evidence for evolution and especially common ancestry, though the evidence (even in the form of dead genes) is at this point somewhat superfluous.

More important, the work tells us what genes may have been useless—and therefore inactivated—in the ancestors of all cetaceans: genes that would be a hindrance to adapting to a fully marine way of life. Now we don’t know that the genes mentioned above were definitely inactivated because of new way of life, but the authors at least provide a suggestive but useful list for further investigation. Are these genes inactivated in all cetaceans? Do they do what they’re said to do when they’re active, and is their inactivation useful in cetaceans? This is all grist for further research.

Finally, what mechanism led to the gene inactivation? The authors posit mutation followed by the generation of “neutral” gene forms, with the inactive forms fixed in cetaceans by genetic drift:

Many of these gene losses were likely neutral, and their loss happened because of relaxed selection to maintain their function.

Well, that might be true, but such fixations of broken genes take a long time, and if they’re neutral they’re likely to keep both active and inactive forms of a gene around for a long time, especially in large populations. And why would the broken gene always be “fixed” in cetaceans rather than coexisting with the active form?

Given that the authors speculate that the loss of gene function might be adaptive in cetaceans, it seems more likely that natural selection swept the broken genes to fixation because they were adaptively superior to active genes (see above for reasons why). It’s a challenge for future work to try to determine, through DNA-sequence analysis, whether broken genes come to be fixed by positive selection or by random genetic drift due to “relaxed selection” (i.e., no selection either way on broken versus non-broken genes). That kind of analysis, as I said, is hard, but it helps us understand how genes get broken when they’re not useful.


Huelsmann, M., N. Hecker, M. S. Springer, J. Gatesy, V. Sharma, and M. Hiller. 2019. Genes lost during the transition from land to water in cetaceans highlight genomic changes associated with aquatic adaptations. Science Advances 5:eaaw6671.


51 thoughts on “How the whale lost its genes

  1. This is very interesting!

    I’m intrigued by the role of temperature and pressure- since these are broken genes, I suppose temperature is not playing a role in selection…?

    What is the body temperature of a whale, and compared to the terrestrial ancestors?

    1. 38 degrees Celsius, or 100 degrees Fahrenheit. Deep ocean water is between 0-3 degrees Celsius (32-37.5 degrees Fahrenheit)!

      1. Isn’t water densest/heaviest at 4°C? Itself quite weird, but shouldn’t we expect a temperature closer to 4°C at the bottom?

        1. You are thinking of pure water Nicolaas. Salt water is most dense at its freezing point, unlike fresh water, which is most dense at about 3.9°C. The coldest LIQUID seawater recorded is minus 2.6°C.

          Below is a temperature/density chart for water with salinities from 0 PSU** up to 45 PSU – ocean water is typically 30 to 35 PSU. As an example, water with a PSU of 40 achieves a maximum density at around minus 5°C [theoretically – I assume it hasn’t been measured outside the lab].

          ** PSU = Practical Salinity Unit

          1. Thanks for the info. I checked Practical Salinity Unit, which is a unit based on the properties of sea water conductivity. “Practical” probably because it is relatively simple to measure at remote locations across and at depth in the oceans.

            1. Exactomont mon petty shoe! On the instrumentation side it is very important to understand what is being measured & to bear it in mind. For example a TDS meter supposedly indicates the total dissolved solids (TDS) of a solution, i.e. the concentration of dissolved solid particles. BUT it can only measure IONIZED dissolved solids, such as salts/minerals – if you chuck sugar or fats [such as milk] into the water a TDS meter doesn’t ‘see’ such particles since there’s no effect on the electrical conductivity.

              While TDS meters are convenient, we don’t know what amount of un-ionized particles are in seawater, nor what effect they might be having. I’m thinking of ocean circulation climate models for example. There’s quite a lot of fat in miniscule sea critters.

              1. You may be referring to me as a cream puff or a cabbage. Either way, I accept the compliment. 😎

              2. May wee, it’s a complan via un fromage [attenday un vidéo en bass] à Derek ‘Del Boy’ Trotter of Nelson Mandela House, Dockside Estate, Peckham, London – CEO/MD of Trotters Independent Traders [T.I.T.]. “This time next year, we’ll be millionaires!”


              3. It’s a fait a complen, then. We shall indeed become millionaires. I don’t know of that TV series would go over so well here. Too many mysterious references. Not like Doc Martin. But, it’s amusing.

      2. so the whale heats itself up because there’s no melatonin – so on land, it’d overheat…. I suppose that’s why beaching is so bad – not just that they are in the direct sunlight, but their hormones are heating them up.

        1. That sounds plausible. Another factor in over heating on land: water conducts heat away from the whales surfaces 25 times faster than air. Without the water as a heat sink, metabolic heating would be lethal.

  2. Cetaceans also are the only obligate carnivores that don’t have a gallbladder. Some herbivores and some omnivores have lost this little organelle,like the hippo, horse and rat, but it is retained in every other mammalian obligate carnivore. Why would a creator design the whale to only eat fish and not give it the ability to digest meat like every other mammalian carnivore on the planet? More evidence that cetaceans evolved from a herbivore who had entered the maritime environment without its gallbladder. To make up for is loss, sperm whales have enlarged the duct that carries bile from the liver to the intestine, to serve as a novel bile storage organ, and is a great example of convergent evolution. It is exactly what you would expect evolution to accomplish and the complete opposite of what a designed organism would look like.

    1. That’s interesting

      I’m confused though – herbivorous whale ancestors lack a gall bladder, whales lack a gallbladder… gallbladders are for digesting “meat” but not fish…

      Does that mean herbivores can eat fish but not meat?

      1. The bile from the liver is an emulsifier to allow efficient fat absorption that presents itself in eating other animals whether fish or other meaty species. Sorry about the confusion of “meat” being different from “fish”. It’s really the fat the the bile is after, something that is not a big concern for herbivores or for animals that continually forage, like the rat, and don’t require a ready reservoir of concentrated bile. Herbivores without a gallbladder could likely digest protein from meat but might have a difficult time handling the fat presented in a Viking buffet around the winter solstice.

        1. That makes sense. Though I still wonder about the diet – there’s still fat in plants, e.g. nuts. I wonder if it’s more complex – perhaps the bile is mostly working on the complex lipids of animals.

          1. It’s likely a matter of quantity. The gallbladder not only stores bile, it concentrates it, which allows carnivores to really digest that burrito bigger than your head. If your only occasionally presented with a low concentration of fat, the liver can make enough bile to directly inject it into the digestive stream and you don’t need the costly and dangerous gallbladder (there are nine ways your gallbladder can kill you. There’s an interesting book, Death by Gallbladder, that details this).

            1. I am pointing to the lipids – you can look them up but I think they will vary across the kingdoms (which I have to look up). A lipid has fatty acid (as you point out), attached to a glycerol-phosphate unit, and can get complex from there with other things I think like peptides, carbohydrates, with other things in the membrane like cholesterol (animals), vitamins, etc… but I’d have to look it up.

              [ … looking it up …]


              Yeah so there’s a lot to read about here for me, at least.

              But the fatty acids have fairly uniform trend in structure and exist in many kingdoms of life as the same structure… that is, I don’t know if any fatty acids that are only found in one kingdom and not others…

        2. I wonder–and you may know–whether the generally less saturated character of marine fat makes large amounts of bile less of a necessity for marine carnivores. Do you know whether this is the case?

          1. I’m not entirely sure that marine sources of fat is less saturated than other terrestrial fats, but they do have more omega-3 fatty acids which have cardiovascular and anti-inflammatory properties. But by and large, waxy saturated fats or unsaturated oils (mostly from plants but can come from animals) are both immiscible with water and require some emulsifier to digest. To digest fats you need bile; to digest a lot of fat at once it’s good to have a gallbladder. The marine cetaceans without one may have adapted by slowing the emptying the stomach, developing a new bile storage organ, like the sperm whales, (see above), concentrating the bile that comes out of the liver (this is what rats do) or not eating so much at once.

  3. Highly readable. I first got interested in the evolution of cetaceans when for amusement I occasionally visited creationist websites just to see what arguments they mounted against evolution. One absurd YouTube posting by David Berlinski was on the “impossibility of evolution of whales” from any land ancestor. He mounted an argument from incredulity…. his hypothesising the start point as a cow and then ending with the whale. His argument he said was “mathematical” in nature… the high number of changes required. I suppose a cow was chosen as a starting point as this seemed to perhaps be most functionally different from the whale. Researching whale evolution to counter his argument in the comments section was a delightful exercise for me…. it being a marvelous example of the progressive nature of adaption and the impact of natural selection.

  4. Well that’s neat. I knew about the olfactory receptors, they’ve been a standard example for a while, but the other classes really add weight to the argument (as if any were needed). Alternatively, perhaps god had a pseudogene synthesizer that he used to test our faith…

  5. I try to stay generally up-to-date on whale evolution because it’s such a great topic for teaching. It’s so much easier to keep up when these detailed summaries are provided!

    I think you mean that they they excluded the genes that looked broken (rather than intact) in the hippopotamus (i.e., they kept only the ones that appeared functional in hippo). Thus they zero in on losses of function that are unique to the lineage leading to whales.

  6. I like how the post shows how many interesting questions crop up from study of this paper, leading to more interesting questions, illustrating how interesting nature is. That creationism is acknowledged here is valuable. It is valuable because of the picture we see by holding that interesting story up alongside the miserable and pedantic circular reasoning of creationism and not-very intelligent design, that we see creationism – indeed, a pseudoscience and not science – leads nowhere except into a mind confused by religion.

    1. Should also note that the broken coagulation genes gives the lie (again) to this system being “irreducibly complex”.

  7. Very interesting, but I note they only looked into disabled genes, and how that can be functional.
    I wonder whether the cetaceans have ‘new’ or ‘changed’ functional genes, though, they surely must have some?

    1. Yes, but those would be harder to find and especially very difficult to find out what they do. Also, a lot of the changes may be due to differences in gene regulation and the sequences of regulatory regions, and those are even harder to pinpoint.

  8. Reading your section in Why Evolution is True on cetaceans was one of the most enjoyable parts of your book. This new research adds another significant and fascinating piece to the cetacean puzzle. Thanks much for bringing this to our attention. You might need to add it to WEIT’s 2nd edition. 😉

  9. Nice. Just to emphasise that we do read and enjoy the science posts, even if we don’t have as much to say about them.

    You would think that mammals in an aquatic environment would be evolving “quicker” than those on land, in the sense that they are adapting to “new” environment. Is there any evidence for this overall?

    Or do land environments change quicker than sea environments, cancelling this out?

  10. Neutral drift versus selection is often going to be hard to work out. For one thing, it is possible that the history of a particular gene includes some of each. We would like it to be cut and dried; nature is not so fussy.

    “And why would the broken gene always be “fixed” in cetaceans rather than coexisting with the active form?”

    This question may not be quite right, as the study involves genes that were fixed. It doesn’t show that there aren’t any genes that could have been broken but by chance are not. I don’t think it shows that there are none that exist as a mix of broken and unbroken either.

  11. Not being a biologist but a mere philosopher with a passing interest in evolution, this is one of the excellent things about WEIT blog. I am kept up to date on issues which I would not normally find out about, such as this paper. So Jerry need not despair if he thinks his science stuff in the blog is not read as much as other stuff. It is!

  12. Thank you Lord, for a science post.

    In these times it makes it worthwhile to log on each morning..

    Seriously, thank you Jerry!

  13. Thanks, much appreciated with the heads up and digest! I had seen the press release title in my feed but had no time to find out just then.

    So, interesting work. On this though:

    The authors posit that most of the broken genes in cetaceans accumulated by neutral processes, but it’s very hard to distinguish that scenario from an increase-by-selection argument, as this involves comparing DNA sequences and looking for a “signature of selection”: nearly impossible in such data. … Given that the authors speculate that the loss of gene function might be adaptive in cetaceans, it seems more likely that natural selection swept the broken genes to fixation because they were adaptively superior to active genes (see above for reasons why).

    Steve Gerrard points out the selection [sic!] bias operating in this paper, which would also apply to the problem of identifying the selective forces for any given selective fixation. But in general there seem to be no problem of identifying the amount of loci which are under drift or purifying respectively adaptive selection [ ]. Note how low population size species such as humans have essentially all non-synonymous fixations under drift while for example Drosophila has about 50 %. IIRC mice, which was used as an example in bioinformatics class, has 95 % so in between.

    And: “the neutral theory is the underlying basis of selection tests.” It is much easier to show specific loci under selection (or not) in population wide genetic data using, say, Tajima’s D. One can for example study how often selective sweeps centered around selected loci drag essentially neutral variation to fixation by chance location in the genome.

Leave a Reply