Heritable differences in dog behavior

It turns out that breed differences in  behavior are highly heritable, which is interesting since it’s impossible to define between-population heritability  ( it is so written) .  But although philosophically impossible to define, it’s apparently fairly easy to calculate.

Gene implicated in these behavioral differences between breeds include candidate domestication genes, genes mapped to phenotypes implicated in domestication, genes that show up in those foxes bred for tameness or aggression, and genes that underwent positive selection in both human evolution and dog domestication – which probably means human domestication and dog domestication.

“For example, breed differences in aggression are associated with multiple genes that have been linked to aggressive behavior in humans. Molecular associations with breed differences in energy include genes previously linked to resting heart rate, daytime rest, and sleep duration in humans. Lastly, breed differences in fear were associated with genes linked with temperament and startle response in humans, and several of the genes implicated in breed differences in trainability have been previously associated with
intelligence and information processing speed in humans.”

Interestingly, genotype accounts for more of the behavioral variation between breeds than it does within breeds.   Like humans, most genetic variation in dogs is within-breed rather than between-breed ( 85%-15% in humans, 70% to 30% in dogs), but that whole-genome number does not matter: what matters is the distribution of the particular genes that influence behavioral traits, not neutral or  behaviorally-irrelevant  genes.

You can’t say any of these things without detailed understanding of the causal mechanisms, just as we could not eat anything until we’d learned that edible sugars are right-handed molecules while amino acids are left-handed.

As for comparing dog breeds and human races, wrong because too obvious.

This entry was posted in Uncategorized. Bookmark the permalink.

65 Responses to Heritable differences in dog behavior

  1. Steve in Greensboro says:

    As the proud owner of a German Shepherd of the German Working Line, I can confirm that in her case at least, she expresses all of the behavioral characteristics of her forebears; with no training at all and no prompting at all she is an intimidating terror to anybody outside her family.

    Just like intelligence, no doubt the genes for aggression in humans will likely be identified in short order, further narrowing the ground on which the race-denialists stand.

    • William H. Stoddard says:

      Let us hope that doesn’t lead to efforts to breed them out entirely, as it did in the future history of Heinlein’s Beyond This Horizon.

    • TB says:

      Just like my neighbor’s new shepherd pup. Nasty creature. Barks at anything not in it’s own back yard. Fortunately, it has calmed a bit in recent months.

    • JP Irwin says:

      I’m worried about them identifying PGRS for ‘troublemakers’. We’re all headed for a FEMA camp I tells ya!

  2. You have done a very good job of explaining why we should not read this already discredited paper, nor draw any unsafe conclusions from the pseudo science contained within it. I hope we can track down the authors and ensure they never speak to any students. Onwards and upwards into 2019.

    • ecgwine says:

      It is becoming harder to spot the sarcasm around here. I can tell you are being sarcastic, of course, but it took my brain two seconds. As for Cochran’s “wrong because too obvious”, I am honestly unsure.

      • amac78 says:

        As for [whether] Cochran’s “wrong because too obvious” [was meant as sarcasm], I am honestly unsure.

        That was indeed a sarcastic tip of the hat to the fashionable wisdom of the Current Year.

        Cochran’s recent (1/1/19) twitter exchanges with the NYT’s Pulitzer-winning Genetics Commissar Amy Harmon provide some context (once you make sense of that kludgy platform). Extracts (some out of sequence):

        @amy_harmon — Yes, David Reich, a Harvard population geneticist, said this. Which makes it all the more salient that, as I wrote, “he explicitly repudiates Dr. Watson’s presumption that such differences would “correspond to longstanding popular stereotypes.’’ That’s why I chose HIM to quote!

        @gcochran99 — [For Reich’s quote to make sense, one would] have to expect that significant genetic influences pushing in a given direction usually result in the opposite effect. Why would that happen?
        Reich is lying, of course, and you prefer that.

        @amy_harmon — Always super-fun to write about race and intelligence because of the comments telling you how dumb you are.

        @gcochran99 — Fishing for a ridiculous quote that confirms what you’d like to believe isn’t necessarily dumb, but it sure is dishonest.

        @gcochran99 — I once talked with a different kind of reporter. We were nattering about IQ, and she said, “Isn’t that a natural consequence of the central limit theorem?” No Harmon she.

        @gcochran99 — If it makes you [Amy Harmon] feel any better, I’ve talked with reporters who were far dumber.

        @gcochran99 — To be fair, it can’t be easy to report on complex subjects that you personally haven’t studied in any detail. [[And that] you haven’t mastered, not even enough to detect glaring errors or falsehoods.] All too often you end up talking to authority figures, some of whom are totally clueless. Like Frances Collins.

        • AppSocRes says:

          Would you please provide a link to the Harmon/Cochran exchange. I expect that it’s a duel of wits where one side entered the arena unarmed.

          Harmon strikes me as someone with a modest gift for writing who was “educated” and later promoted beyond her intellectual capacity. I learned from Wikipedia that Harmon majored in American Studies at UM, Ann Arbor. When I was at Brown in the 1970s – before all the other x-studies departments popped up like mushrooms – we used to joke that American Studies was for those who wanted to major in American Lit, American History, etc., but didn’t quite have what it took.

          • MEH 0910 says:

            • Jacob says:

              Re: central limit theorem, was she suggesting that each person getting a random sampling of +/- alleles should make the curve normally shaped?

              Smart, but nonrandom mating should fatten up the tails. Something like a t distribution. But that’s actually how the “bell curve” looks, I’m pretty sure.

              Someone who knows more math than I do could find out whether the curve we actually get is meaningfully different from a bell curve that got flattened by assortative mating: ditto for personality traits, height, BMI, take your pick.

        • ecgwine says:

          Interesting, this. Not Harmon, of course, who is a journalist and as such almost professionally obligated have nothing interesting to say on her chosen topic. But I just checked out Reich’s quote. I understand Reich is doing a lot of valuable and groundbreaking work on exactly the type of questions that have bearing here. And I can only assume that he is keenly aware that he is walking a political tightrope (he must know he is never further than a slip of the tongue away from professional ostracism), so he may be excused to eagerly proclaim agnosticism what is merely plausible — but his active use of the tired invective of “racist” against those who do argue that the plausible is … plausible does smack of cowardice.

    • Jacob says:


      Man, what gets these people thinking they can just say that?

      I want to know where the Discreditation Fairy lives. Do I have to go spelunking in Jim Flynn’s rectum to meet this guy and ask him why he only discredits the most useful and predictive ideas?

  3. mapman says:

    It turns out that breed differences in behavior are highly heritable

    There was never a poodle who behaved like a typical jack russell terrier. So far most high tech genomics merely confirms common knowledge and stereotypes.. Hopefully that is just a learning phase on the way to big discoveries.

  4. albatross says:

    Raising English Setter puppies is pretty instructive for this kind of thing–after the first time you watch a puppy with no training at all running around in the yard, “pointing” every bee and butterfly it sees, you’ve got a sense of how much genes can matter….

    • ANONYMOUS says:

      I sometimes ride mountain bikes with a man who brings his blue heeler on slower group rides. The dog is generations removed from any trained working animals.
      It still runs behind and nips at stragglers and runs along the edges of the group nipping at anybody who veers off trail.
      It’s such a very specific behavior that it makes you wonder if maybe we are missing something. Can such a complex behavior really be transmitted along generations just with base pairs and protein folding?

      • gcochran9 says:

        “Can such a complex behavior really be transmitted along generations just with base pairs and protein folding?”


        • Patrick Boyle says:

          “Can such a complex behavior really be transmitted along generations just with base pairs and protein folding?

          I’ll tell you something even more improbable. Look at an old fashion LP record. look close see that squiggly line? Yes, but can you see Mozart’s music in that groove.

          The answer is that it’s just coding. Almost anything can be encoded. Burn the right little pits on a disc and you have ‘Saving Private Ryan”.

          • gcochran9 says:

            There’s a guy that can identify the composer by looking sideways at the grooves in a vinyl record.

            • ziel says:

              Believable. I’m sure I could tell a Mozart symphony from a Tchaikovsky symphony that way – WAM’s would have very predictable groove densities by movement as well as in movement, while PIT’s would have much more dynamic changes throughout.

              Distinguising Mozart from Haydn would be a helluva lot more difficult, but surely possible with training.

              • Ziel says:

                Of course since Mozart symphonies fit 4 movements to side while Tchaikovsky only 2 per side that’s a no brainier. But I’d imagine you could tell one Beethoven symphony from another by looking at groove density.

      • dearieme says:

        Long ago, briefly, tangentially, superficially, I worked on protein folding. Dear Christ, what a problem! End of complaint.

      • Greying Wanderer says:

        million random behavioral patterns

        select the one that by chance happens to be useful.

  5. Smithie says:

    I have known retrievers that would hardly twitch, if you accidentally stepped on them. Another dog, a pitbull mix, would nearly leap at your throat, if you stepped too close to him, let alone on him. He was put on some sort of meds.

  6. Eponymous says:

    “70% to 20% in dogs”

    I assume you mean 70/30? So twice as much between-group variation as in humans?

    Incidentally, do you know why it’s a good deal higher in dogs? My understanding is that FST is driven by neutral variation, so selection wouldn’t affect this much. Is it just a matter of generations since breeds diverged and degree of gene flow? Or small founder populations of breeds?

    • ecgwine says:

      I think it stands to reason that dog breeds will be more inbred than human populations. I would have expected the discrepancy to be even larger.
      This is also why I think that while human races and dog breeds do indeed offer themselves as obviously analogous comparanda in principle, you can hardly expect the two to be extremely similar.

      • Eponymous says:

        Can you take me through the math? How many generations ago did dog breeds diverge compared to human racial groups? How much gene flow has there been since?

        • ecgwine says:

          I haven’t done the math, this was just my naive expectation. I am not an expert on this and I am sure many people here can answer this better than I can.
          However, I just googled this, https://doi.org/10.1093/jhered/esg004 — and I am surprised to find that fixation indices (FST) between dog breeds was (for the loci considered in this study) at 23%, which is surprisingly similar to FST between human races, so maybe the analogy is closer than I expected.

        • gcochran9 says:

          MAny breeds are very new, ~150 years.

    • gcochran9 says:

      Yes, 70/30. Thanks, fixed.

  7. Rosenmops says:

    We had two pugs from the same litter. They were with us for about 15 years. There personalities were very different. Dogs are definitely born with personalities, just like humans.

  8. Philip Neal says:

    I know nothing of these matters, but I was intrigued by the bit about genes that have the same phenotypical effect in dogs and humans. Does this mean that it will one day be possible to quantify resemblances between different sorts of humans and different sorts of dogs? The distance between Tony Blair’s genome and the poodle genome, say?

    • ghazisiz says:

      “The distance between Tony Blair’s genome and the poodle genome, say?”

      Sounds like a business idea. Upload your DNA, and for a fee we will calculate the distance between you and the various dog breeds. Might pitch this as a way to help you find the most compatible dog breed. A good marketing campaign might be to identify the breed closest to various celebrities (Tony Blair, Obama, Madonna, etc.). It wouldn’t matter if the whole thing is scientifically bogus, because that never matters in business.

    • Greying Wanderer says:

      “genes that have the same phenotypical effect in dogs and humans”

      i bet a lot of cops have the same combo of aggression+restraint genes that sheepdogs have

  9. Xenophon Hendrix says:

    So, which human populations are the most domesticated and which are the least?

  10. Re: Reich and the tightrope. I am rereading “The Horse, The Wheel, and Language” now that I am much more sensitised to the political arguments. Right at the outset, Anthony makes an impassioned rejection of some arguments suggesting that there might be races and they might be meaningful (pg 12). He assures us that the lack of absolute boundaries for races, the difficulty in defining what a race is even among those who supposedly agree about it, and the horrible history (Germans! WWII! Eugenics! White Supremacists!) are insuperable obstacles to there being anything about race in the discussion of populations. On the very next page (14, after an illustration), he dismisses the need for such rigor in arguing that language families actually do exist. I have encountered this sort of incantation several times “We wouldn’t want you to think we’re racists. No, no, none of that here. We reject that entirely.” always right at the outset, to put critics at ease. (Except I doubt it does.) As I don’t know these individuals or their editors, all writing very similar things in order to get their semi-controversial ideas heard at all, I don’t know if they have convinced themselves that these things are true or are just trying to protect themselves. Are they going as far to the edge as they dare in hopes of keeping a door ajar for those who pick up the cues, or are they hall monitors making sure we don’t even enter the Senior Wing without a pass?

    • ecgwine says:

      I think it varies between authors. This kind of propitiation is particularly crass in “Guns, Germs and Steel”, right at the beginning of the preface, Diamond goes “The question motivating the book is: Why did history unfold differently on different continents? In case this question immediately makes you shudder at the thought that you are about to read a racist treatise, you aren’t” — this is quite impressive for 1997, to “shudder” from visceral anti-racism merely by the mention of there being differences between continents. Of course in this case, this passage is almost naively betraying the author’s ideological biases.

  11. Ilya says:

    Dr. Cochran: thank you for publicly, via twitter, taking a convincing stand for truth and Dr. Watson, under your real name. I didn’t see many people doing this.

  12. Crew says:

    Well, since dogs and humans cannot interbreed (despite valiant attempts by some), heritability of dog traits has no impact on humans, and Lewontin wins again!


  13. Zeinish says:

    Dear Mr.Cochran, we all know what you are against, but what are you for? What do you want? What is your plan? What is your endgame?

    What you think about Departments of Eugenics, on state and federal level? What you think about Ministry of Eugenics in every country of the world?
    What you think about United Nations Frameworks Convention on Eugenics, Intergovenmental Panel on Eugenics, High Level Advisory Group on Eugenics, International Council on Eugenics, United Nations Eugenics Task Force?

    If you win, if you succeded in your life work and make HBD-IQ genetic science accepted by all educated people worldwide, this is exactly what you will get.
    You better to like all these things.

    • gcochran9 says:

      I can only hope the your day job does not involve prognostication.

      • Zeinish says:

        What is your prediction, then? What will happen when you life work succeeds and averyone is persuaded to accept the truth about HBD and IQ?

        • Rye says:

          A generally accepted tenet of the European intellectual tradition is that policy based on fact and evidence is more likely of achieve our desired goals than policy based on blind faith and lies. If we allow our societies to become structured around lies then we should not be surprised when we come to be ruled by professional liars and opportunistic parasites. If we minimize the number of absurd lies which our policy makers are forced to kowtow to, we might actually get some honest and principled leaders who will fight for our interests.

    • It is true that accurate information can be misused. Guns are owned by some bad people. Hell, nuclear weaponry is owned by some bad people. Things can go wrong. However, there is a strong tendency for things to go even worse with inaccurate information.

      • ecgwine says:

        The question “what is your endgame” asked of a scientist only betrays how far gone the questioner is ideologically… If “Zeinish” was at all being genuine, I find this frightening.

        If “HBD-IQ science” corresponds to reality, we have to study it in order to be better able to cope with reality. If it isn’t, well, just let it be thrown out (or incrementally corrected) by the normal scientific method, not by moralising about it.
        The implication that the “educated” should consciously steer away from what is factual because some facts may be unwelcome, or that you are required to “like” the facts you report (so, climatologists need to “like” their projections before they are permitted to publish them?) — such a mindset strikes me as deeply perverse and unenlightened to say the least.

      • dearieme says:

        “Guns are owned by some bad people. Hell, nuclear weaponry is owned by some bad people.” The best thing I’ve heard Trump say was “You think our country’s so innocent?”

        • Zeinish says:

          Do you want all details of nuclear, chemical and biological weapons design and production be declassified? Information wants to be free, after all.

          • Texan99 says:

            Do you think that true and easily obtainable information about heredity and IQ can be hidden from the unwashed public in the same way that some sophisticated weapons technology can temporarily be hidden or monopolized? There’s a big difference between keeping a dangerous secret and suppressing honest public debate among people who are already quite aware of the facts.

    • JRM says:

      The goal is colonization of the galaxy. Read the Manifest Destiny post.

      • Zeinish says:

        Squishy sacks of rotting meat colonizing galaxy?
        If Galaxy is the goal, concentrate on AI research and skip everything else.

        The first explorers of Earth had long since come to the limits of flesh and blood; as soon as their machines were better than their bodies, it was time to move. First their brains, and then their thoughts alone, they transferred into shining new homes of metal and of plastic.
        In these, they roamed among the stars. They no longer built spaceships. They were spaceships.
        But the age of the Machine-entities swiftly passed. In their ceaseless experimenting, they had learned to store knowledge in the structure of space itself, and to preserve their thoughts for eternity in frozen lattices of light. They could become creatures of radiation, free at last from the tyranny of matter.

    • another fred says:

      Zeinish’s post is instructive of a mindset that is evident in a large portion of the population. Having “government” (at some level) in charge of maintaining whatever state of social homeostasis is possible (or desirable in their minds) is their default position.

      The idea of free people in a dynamic society is unimaginable. The bullies/Nazis will take over!

    • Tanturn says:

      Could you explain exactly what wpuld be the bad policies you fear?

      • Zeinish says:

        What happened the last time when “race science” flourished and fear of “hereditary degeneracy” was rampant?

        • Tanturn says:

          Sane immigration policies? Less fetal alcohol syndrome? Yes, there were some abuses committed in the course of forced sterilization, but that’s no reason to throw the baby out with the bathwater. “Reproductive rights” are never absolute. Our government already takes away children from mothers too retarded to raise them, and tries to prevent them from having sex. What’s wrong with forcibly preventing those same women from getting pregnant? Similarly, if our government has decided that someone is such a danger to society he should be locked in a cage, why is it such a horrible violation of his “rights” to forcibly sterilize him?

          Eugenics had little to do with hatred of the Jews. Hatred of Jews existed long before it, and continued to exist after its “discrediting.”

        • ecgwine says:

          I guess they just close the studbook.
          When was the “last time” this happened? Not an expert, I believe this happens fairly often (1990 for the Freiberger breed). You can always re-open it if it turns out the closing was premature.

        • Rye says:

          What happened the last time when “race science” flourished and fear of “hereditary degeneracy” was rampant?
          Europeans had a thriving civilization which ruled the world?

        • magusjanus says:

          Well the last couple times someone took “equality” seriously enough we got Khmer Rouge wiping out 1/4th of Cambodia and Mao starving 30-40mio people to death. Less dramatically Venezuela has squandered a few decades of oil and turn its currency into a joke and its people to eating zoo animals.

          So maybe you’re onto something and we socially ostracize if not outright persecute any discussion of ‘equality’ lest some pinko take power and ruin the country.

  14. Efim Hawkins says:

    One of your all-time best posts, in my opinion.

  15. Robert A Gressis says:

    View at Medium.com

    Does anyone have any response to this? I can’t follow the math, because, though I am high IQ, I am a pretty stupid person.

    • Ursiform says:

      The math has essentially nothing to do with his arguments. His point pretty much seems to be: If a correlation isn’t perfect and if it leads to a conclusion I don’t like then I can say it isn’t real.

    • RCB says:

      A “valid measure”, whatever that means, should correlate just as strongly in the tails as in the center… because… I said so. Blah blah blah. I’m going to mention fat tails because I’m Taleb and that is pretty much my whole career. Blah blah. IQ, devised to measure intelligence only, doesn’t ALSO measure entrepreneurialism, savvy, moxie, and other things that matter in life… therefore it’s invalid. Blah blah. Problematic. Blah.

    • Greying Wanderer says:

      to be “useful” a measure should correlate with the center first as that’s where the majority of people are and for a measure to be “valid” it only needs to be useful. if it doesn’t fully correlate with the tails then let people research why.

      say for example there was AQ (athletic quotient) made up of the average of four 0-10 scores in each of running, jumping, climbing and swimming and a person scored 10 on the the first three but zero at swimming then their AQ would be 7.5 which would underestimate their ability at the first three and massively overestimate their ability at swimming

      but most people wouldn’t be like that so AQ would be a useful measure for most people.

      maybe IQ is like that: multiple components averaged together and for most people it mostly works with some people or populations who either have particularly high or particular low scores in particular traits which skew their average.

Leave a Reply

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

WordPress.com Logo

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

Google photo

You are commenting using your Google 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