Ketosis as a way of life

There’s a new article out in ASHG that discusses a regional selective sweep in CPT1A, carnitine palmitoyltransferase 1A, which plays an essential role in fatty acid metabolism. A mutation has become extremely common, with a gene frequency > 50%, in northeast Siberian populations, Eskimos, and Aleuts.  This happened even though the c.1436C>T mutation has some negative side effects, such as reduced fasting tolerance and, apparently, a higher risk of infant death.

It looks as if this variant conferred significantly higher fitness (~15%) over the past few thousand years and has increased very rapidly.  Almost certainly, this has something to do with traditional diets in these areas, in which fat provided 80-85% of energy and protein 15-20%, with hardly any carbohydrates. We’ll know more about the biochemical details of how this mutation works when someone makes a mouse model.

This could only happen in the Arctic, where agriculture is impossible, plant foods limited, and marine mammals use blubber as insulation.

It illustrates an important point, mentioned in our book: an ancestral diet might have some payoff, but likely the appropriate diet would be the foods eaten by your ancestors, not somebody else’s.

This entry was posted in Dietary adaptations, Eskimo, Genetics and tagged . Bookmark the permalink.

67 Responses to Ketosis as a way of life

  1. So, in physical and probably psychological terms, were are getting the benefits of choices made by our ancestors, with at least a 4000 year lag? The ultimate marshmallow test.

  2. Fourth doorman of the apocalypse says:

    Sorta, kinda, OT (off topic) and OT (on topic):

    Once of the defenses against influenza is mucus, as described here:

    Could it be that having one copy of the cystic fibrosis mutation was part of an early defense against influenza?

  3. Rosemary: Browsing Shakespearean era cookery awhile back I was a bit surprised at the prevalence of rosemary as an herb. For whatever it’s worth and doing a little browsing just now — searching on the term rosemary at GodeCookery yields 36 hits. (Rosemary, by coincidence of course of course, has always been one of my favorite herbs.) ..

    • gwern says:

      And sage also seems to have fallen by the wayside. I find that too bad as sage & rosemary are two of my favorite herbs.

      • Sage pollen happens to be a common allergen. For the local pollen report here in New Mexico the pollen count for it is usually grouped with grasses.

        • karenjo12 says:

          I believe that kind of sage is different from the culinary herb. And where do you all live that people don’t cook with sage and rosemary? I want to avoid that place.

      • melendwyr says:

        Strange, given how popular it was in medieval cookery – as medicine rather than for its culinary properties, although they tend to blend. It’s named after a word meaning ‘wisdom’ for a reason. Tisanes made with it are delightful – and if taken in sufficient quantities on an empty stomach, can cause dizziness and hallucinations. Has a higher thujone content than absinthe.

    • As Ophelia says, “There’s rosemary, that’s for remembrance.”

  4. j says:

    Sir, no more Dr Atkins diet for me. I’m going back to Egyptian style fleshpots and eating bread to the full.

  5. dearieme says:

    (i) the prevalence of rosemary: it’s one of the Mediterranean herbs that grows very well in England.

    (ii) I remember reading that the Eskimos got part of the vegetable part of their diet from the stomachs of walruses and whatnot, albeit somewhat second hand. I imagine I learnt that at the age when small boys were expected to know that sort of thing lest they ever found themselves having to survive in the Arctic.

  6. mischling says:

    What if you are half Ashkenazi half northeast Euro what would be the best diet then?? O_o

  7. Whyvert says:

    “the appropriate diet would be the foods eaten by your ancestors, not somebody else’s”

    Foods eaten by one’s ancestors is a moving target. Lots of current staples have changed a lot recently by selective breeding. My ancestors’ wheat isn’t the same as my wheat.

    • Fourth doorman of the apocalypse says:

      OK, but in what way has it changed that reduces the ability of your genes to deal with it?

      • MawBTS says:

        Most wheat today is high-yield semidwarf wheat, created in 1960 by crossing a Norin strain with Brevor. It was bred with a “maximum yield über alles” goal in mind…who knows if it’s actually that good for you.

        They’ve found pots of honey in Egyptian tombs that are still edible after thousands of years (the honey, that is, not the pots). I could have sworn I once read about ancient Egyptian grains being used to make bread by an archaeologist, but I can’t find anything about that now. I wonder how different it would taste to today’s bread.

        “Mmmm…Just like great great great (…) grandma used to make”

  8. Uptown Resident says:

    23andMe should put some R&D into giving clients customized diets based on their ancestry reports.

    That would be an easy way to mainstream their product and also get a piece of the massively lucrative diet and wellness industry.

    • MawBTS says:

      Cool idea, but that would fall under the scope of health-related advice and they’re in enough trouble with the FDA already.

      • I know, but I heard they’re fighting the FDA’s ban. It seems like they got in trouble mainly because they initially blew the FDA off, and could have kept the health advice had they gone through all the hoops.

        My husband and I had our kits done last Christmas, downloaded our raw data, and ran it through Promethease’s software, for $5, which did the same health analysis that 23andMe used to provide.

  9. Tore says:

    Looking forward to more ancient coprolite dna results which can show old diets and what kind of bacteria that thrived in different peoples diets. The gut-brain connection is gaining momentum now I believe:

  10. MawBTS says:

    The Mongols would have had a close-to-ketosis diet, wouldn’t they? Though they drank milk, which has carbs.

    Offhand, does anyone know the macronutrient breakdown of horse blood?

  11. Josh says:


    I was reading that an animal’s ability to recognize its own reflection in a mirror is a sign of intelligence. They have a test for this.

    I was wondering if anyone has tried doing this with toddlers or babies of different races?

  12. hbd chick says:

    milk! whole milk. raw milk. butter, too. lactase persistence, ftw! (^_^)

    (but i’m not giving up my chocolate, d*mnit!)

  13. Matt says:

    It illustrates an important point, mentioned in our book: an ancestral diet might have some payoff, but likely the appropriate diet would be the foods eaten by your ancestors, not somebody else’s.

    This has less practical application than you think though. Which ancestors? How long ago? Etc.

    Potatoes are a major staple in European cuisine now. What should do with them? What about tomatoes? Peanuts, coffee, tea, etc?

    What about grains more generally? And milk? What about the ancestors before grain and milk consumption?

    • JayMan says:

      Think adaptation time.

      The milk example makes it most clear.

      • Toddy Cat says:

        I vaguely recall an interview that Cochran gave over at the old “2 Blowhards” site a long time ago, where he said that the people who were most likely to benefit from an ancestral diet were late agriculture adapters, like Blacks, certain Indians, and Aboriginals, and that unfortunately they were probably the people least likely to adopt it. Sadly probably true.

      • Matt says:

        Yes, but “adaptation time” relative to what? We’ve been vertebrates far longer than we’ve been farmers and milk drinkers. “Time” itself doesn’t really tell us that much, especially since rates of evolution vary across time intervals. And ultimately the time claim is premised on specific physiological features, not time itself. Milk is supposed to be good for us not because we’ve been drinking it for X years, but because we can break down lactase. But once we start looking at physiology, it’s less clear how healthy certain foods are. Milk for example is more than just lactase. Milk may be loaded with insulinlike growth factors which may stimulate cancer growth for example.

        • JayMan says:


          After enough generations (1000-2000 years worth definitely), a people become perfectly adapted to whatever it is they eat (during that time).

          Don’t buy those claims that X causes cancer.

          • gcochran9 says:

            ” Perfectly adapted” in 1000-2000 years? I wouldn’t bet on that.

          • Matt says:

            I don’t think you understood my point. Time is just a counter here. Milk ultimately is being prescribed not because people have been drinking it for X years, but because we can break down lactase. Presumably, if we couldn’t break down lactase, it wouldn’t be prescribed even if we had been drinking it for X years. So if these prescriptions are ultimately based on physiological mechanisms, as they appear to be, then it becomes more complicated.

            I never said anything about X causing cancer. Insulin and glucose, which lactase is broken down into, have been shown to promote cancer proliferation and chemoresistance in human cell cultures.

          • Ursiform says:

            Pandas remain poorly adapted to eat bamboo.

          • Matt says:

            Yes good point regarding pandas. Some of their traits such as the false thumb which is used to grasp and hold bamboo appear to have evolved to adapt to bamboo consumption. But other aspects of their physiology suggest that they’re poorly adapted to bamboo. They’ve been eating bamboo for a long time, but this length of time itself doesn’t tell us that bamboo is optimal for them.

        • Richard Sharpe says:

          Milk is supposed to be good for us not because we’ve been drinking it for X years, but because we can break down lactase.

          The details actually matter, otherwise you sound like an idiot.

          Take it from me, I have sounded like an idiot before because I was careless with the details, but I learned.

          • Matt says:

            So it’s lactose which is broken down by lactase.

            The details do matter, but I don’t care about sounding like an idiot. And this detail does not alter the issue I raised. I’m not confidently making prescriptive claims here. I’m asking questions to people who are.

  14. Greying Wanderer says:


    “This has less practical application than you think though. Which ancestors? How long ago? Etc.”

    I think the logic of it goes something like this.

    1) assume every food stuff can have a benefit and a cost so the three categories are:

    a) all good no bad
    b) some good some bad
    c) all bad no good.

    Ignoring (c) there’s an evolutionary incentive with each food source to both maximize the benefit part and minimize the cost part (if it exists).

    2) Time is one factor however I think the kicker is likely to be the criticality of that food source at some moment in time and the length of time it was critical so

    evolutionary incentive = criticality * time.

    Hence blubber with eskimo and (imo) milk with NW Euros and (also imo) likely thousands more of varying degrees of magnitude – I think there’s likely to be thousands of little ones which may not be that relevant to diet but might be relevant to medical drugs – partly simply because of eating the same thing over time but I think the big ones would stem from a moment when that food source was critical.

    So foods your ancestors ate that were at some point in time critical to their survival.

    A lot of people will have a jumble of different ancestral food genes but if the ones for Swedes, Japanese, Bengalis etc are identified then people should eventually be able to get a list telling htem they have:

    swedish fish genes
    japanese meat genes
    bengali carb genes
    or whatever

    (although i think a lot of these adaptations may be at a regional rather than national scale)

    • Greying Wanderer says:

      For example a population that at some point in time had only one source of vitamin C and that source had some negative side effect then that population might develop genes that would maximize vitamin C uptake and reduce the side effect.

      In a modern population with a lot of sources of vitamin C those genes might not matter much from a dietary point of view (maybe higher risk of ODing on what is considered a normal intake of vitamin C) but those genes might effect reactions to medical drugs.

  15. Greying Wanderer says:


    “Pandas remain poorly adapted to eat bamboo.”

    That could be an example of the same thing in reverse. If pandas were pushed out of their natural habitat into bamboo their lack of reproduction could be a product of not getting their original diet. Have any zoos fed them on meat?

    (got that idea from someone on here)

    • Matt says:

      Pandas have been eating bamboo for a long time. They’ve been seen eating small amounts of meat in the wild sometimes. I know zoos will give them fruits, biscuits, and ice cream. I don’t know about meat though. But their diets are mainly bamboo and they seem to prefer it. And they seem to have certain adaptations such as false thumbs and they harbor certain gut bacteria that help them break down cellulose. On the other hand, they have carnivore digestive systems. Would they be better off and healthier eating meat? By the same token, we have been drinking milk for a long time and enjoy doing so and have certain adaptations for it, but is it optimal for health?

      • Greying Wanderer says:

        i guess they’d have to experiment with whatever it is they think pandas ate before they were pushed out of their original habitat. It may not be meat they’re missing but some specific vitamin or something – but it would be an interesting experiment to find out.

        • Matt says:

          Yes but if bamboo isn’t that good for pandas, how do we know milk is good for us?

          Here’s a recent study that found higher milk consumption was associated with higher mortality:

          “A study published in the Oct. 28 issue of The British Medical Journal suggests that consuming three glasses of milk per day may double women’s risk of dying in 20 years, compared to drinking less than one glass daily, Medical News Today reported.

          Researchers in Sweden found that the sugar D-galactose— which comprises half of lactose— increases oxidative stress and inflammation of the body. Experimental evidence in various animal species shows that chronic exposure to galactose can expedite aging and decrease lifespan.”

          • JayMan says:


            “A study published in the Oct. 28 issue of The British Medical Journal suggests that consuming three glasses of milk per day may double women’s risk of dying in 20 years, compared to drinking less than one glass daily, Medical News Today reported.”

            That’s a correlational study. It’s worth less than the paper its printed on. Therein stems the confusion that surrounds medical/health advice.

            For the record, most cancers are both not very heritable and are uncorrelated with IQ. This severely limits the role behaviors can have in their genesis (since these are heritable).

          • Matt says:

            Why is it worthless? Observing correlations may not be the final word but it’s basic to induction. The researchers also point to experiments showing chronic infllammation from galactose.

          • JayMan says:


            “Why is it worthless?”

            Reread my comment, and think about it.

            “Observing correlations may not be the final word but it’s basic to induction.”

            We are awash in those at this point (if this was 60 years ago, then sure). As far as health and lifestyle goes, they largely turn out to be uninformative.

          • Matt says:

            Still not sure why observing and noting correlations is worthless. Can you explain? I’m not sure why just because there are already lots of observations, we should stop making them. What qualifies as enough observations? Can you ever have enough?

            Why are they uninformative, especially if they’re consistent with experiments showing chronic inflammation for example?

          • JayMan says:


            Observational studies don’t tell you what the causes of the observed associations are. No matter what fancy chicanery the researchers involved use, there’s no way around that (well, almost).

            Physiological studies are also of limited use. In this particular example, it’s unlikely one has anything to do with the other. Often researchers observe a biological process but are not necessarily sure how what they’re seeing is occurring. They often confuse description for explanation.

            Much of the health wisdom gleamed from observational studies has died when subjected to randomized controlled trials. Genetic controls (e.g., twins, adoptees) are one way of getting at causation. Controlling for heritable characteristics (like IQ) are another. The current medical establishment suffers from this defective research and a failure to take genetic factors into account – among other things.

          • Matt says:

            I don’t see why that would make observational studies worthless. Observing correlations, whether in ordinary sense experience or in observational studies, seems important. Otherwise, where would you start?

            Why are physiological experiments of limited use? What about drug trials? Giving people drugs and seeing what happens, for example, seems like a good way of finding out what they do to people physiologically. I don’t see why explanation is more important than description. In the more rigorous and successful science of physics, Maxwell’s equations mathematically describe electromagnetic behavior and make correct predictions. They don’t really explain anything about what is happening. They model what happens when you do experiments.

            If observational studies can point to randomized controlled experiments to be done, then presumably they do have value.

  16. Anonymous says:

    Pandas are perhaps an example of what happens with extreme niche specialization when the specialized plant prey (bamboo) is winning the co-evolutionary arms race. Humans are perhaps on the opposite extreme with omnivory amplified through agriculture, plant breeding, and food markets and transportation.

    • Greying Wanderer says:

      I think it depends on whether pandas moved into bamboo forest out of choice or they were pushed into because it was the only terrain not suitable for farming.

      I’m assuming it was the latter (don’t know for sure) in which case what did they used to eat? Figure that out and feed them it and I bet they start ****ing like crazy.

      • Matt says:

        I’m not sure if choice would be relevant. They’ve been eating bamboo for a long time and have some adaptations for it.

        Their population decline has to do with habitat loss, not their bamboo diet. Their habitats are lush with lots of bamboo. Large carnivores require more land for survival. If pandas were carnivores, they probably would be even fewer in number.

        • Greying Wanderer says:

          “Their population decline has to do with habitat loss, not their bamboo diet. Their habitats are lush with lots of bamboo.”

          I’m not talking about their population decline in the wild – as you say x pandas per y square kilometres is limited by y – I’m talking about zoos trying to breed them.

          It would be in their interest to experiment with their diet imo.

          • Matt says:

            Yes I agree it would be interesting although I think it could be dangerous trying to fiddle too much with them. It might be better to just try to conserve them as much as possible in their present form.

  17. Greying Wanderer says:


    “Yes but if bamboo isn’t that good for pandas, how do we know milk is good for us?”

    We don’t. We know it was good for some people’s ancestors in a particular region at some time in the past relative to what else was available in that region at that time.

    Like malaria resistance things that protect against one very bad thing – like starvation for example – may have less critical side effects.


    “Here’s a recent study that found higher milk consumption was associated with higher mortality:”

    The trouble with that is the people drinking more milk might be those best adapted to drinking milk but at the same time – and for exactly the same reason – are also less adapted to something else in their modern diet (like sugar for example).

    So it might be true or partly true or it might be a correlation with a group of people who aren’t adapted for something else.

    It’s a bit like the thing many years ago where it was said people who drank a lot of coffee had a higher risk of lung cancer but then it turned out people who drank a lot of coffee were disproportionately heavy smokers.

    So for example did those heavy milk drinkers also have more alcohol related deaths and/or obesity/diabetes related deaths?

    • Matt says:

      Milk does have a lot of sugar though. About 40% of it is sugar. So it could be something else besides sugar. And if milk is good for us, it could mean that sugar isn’t really bad for us.

      • Greying Wanderer says:

        Lactose yes, but there’s forms of sugar that have to be processed by the liver and some that don’t and lactose doesn’t

        or a more sideways possibility,

        if LP also led to genes that gave people a sweeter tooth that may have been a good thing when the only source of sugar for those people was in milk but is a problem now we have HFCS coming out of our ears.

        If so it might also relate to liking alcohol a bit too much

        “A new study shows that how people perceive and taste alcohol varies as a result of genetics. The scientists focused on three chemosensory genes—two bitter-taste receptor genes known as TAS2R13 and TAS2R38 and a burn receptor gene, TRPV1. The research is the first to consider whether variation in the burn receptor gene might influence alcohol sensations, which has not previously been linked to alcohol consumption.” – alcohol tastes gooooood!

        (h/t hbdchick)

        Anyway I’m not suggesting it as an alternative just an example of the idea that the correlation with drinking milk might be a correlation with LP which might be a correlation with not being adapted to something in the modern diet which other people who spent longer under agriculture are better adapted to.

        (If it was an alternative I’d suggest maybe agriculture causes people to lose some of their sweet taste buds or something like that so both more recent pastoralists and HGs have a sweeter tooth which has become a problem now because of fast foods etc.)

  18. Pingback: Eskimo Evolution | West Hunter

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s