Variation in dog coats

by Matthew Cobb

This blog normally (and rightly) gives pride of place to cats, but this post will be about dogs. I trust Jerry will forgive me when he returns.

Hidden away in the same issue of Science that deals with Ardi, there’s an article by a host of researchers from the US and France (you’ll need a subscription to get past the abstract), looking at the genetics of dog coat variation. Amazingly, it turns out that most of the massive range in coat phenotypes we can see between “pure breeds” and mongrels can be accounted for by mutations in just three genes. (If I were one of the authors, I’d be pretty peeved – all the media attention will be focused on Ardi; in a “normal” week, they could be sure of making the TV, press and radio…)

The authors studied three characteristics of the canine coat: (i) the presence or absence of “furnishings” (growth pattern marked by a moustache and eyebrows seenin wire-haired dogs); (ii) hair length; and (iii) the presence or absence of curl. To find the genetic bases of these characters, they created three genome-wide single-nucleotide polymorphism (SNP) data sets, based on a) 96 dachshunds showing three coat varieties (wire-haired with furnishings, smooth, and long-haired without furnishings); b) 76 Portuguese water dogs showing the curl phenotype and c)  903 dogs from 80 breeds representing a wide variety of phenotypes.

They found that variability in three genes, coding for R-spondin–2, fibroblast growth factor–5, and keratin-71, accounts for most of the different kind of coats that can be observed (“these three mutations in various combinations explain the observed pelage phenotype of 95% of dogs sampled”) . They summarise their findings in the following figure:


Interestingly, they conclude “None of the mutations we observed were found in three gray wolves or the short-haired dogs, indicating that short-haired dogs carry the ancestral alleles (table S1). Our finding of identical haplotypes surrounding the variants in all dogs displaying the same coat type suggests that a single mutation occurred for each trait and was transferred multiple times to different breeds through hybridization.”

However, although they open their article by pointing out that dogs and humans have lived together for around 15,000 years, they point out in their conclusion that dog “breeds” are an incredibly recent invention – less than 200 years. Inother words, in a couple of hundred year, artificial selection on just three genes has produced an incredibly variety of phenotypes. Furthermore, that selection may have focused on different characters, such as aggressivity or size, the genes for which may be linked to the coat genes. They conclude, “Consequently, in domesticated species, the appearance of phenotypic complexity can be created through combinations of genes of major effect, providing a pathway for rapid evolution that is unparalleled in natural systems.”

This, of course, was also Darwin’s insight (although he did not put it in these terms). Artificial selection by 19th century breeders provided him with the key to understanding natural selection as the force that could generate apparently directed evolutionary change. And that is a much more powerful explanation than anything the creationists can come up with.


  1. Hempenstein
    Posted October 2, 2009 at 8:17 am | Permalink

    Cool research of course, but also, THANKS FOR POSTING SOMETHING ABOUT DOGS!!!!

  2. Doc Bill
    Posted October 2, 2009 at 8:20 am | Permalink

    “Pride” to “cats?”

    Srsly, Jerry, leave the humor to me.

    • Sili
      Posted October 2, 2009 at 9:35 am | Permalink


      “by Matthew Cobb”

      • Doc Bill
        Posted October 2, 2009 at 9:44 am | Permalink


        “Pride” to “cats?”

        Srsly, Matthew, leave the humor to me!

        (Sorry I barked up the wrong tree.)

  3. Posted October 2, 2009 at 8:27 am | Permalink

    Wow. The Bichon Frise is really, really mutated.

    • Sili
      Posted October 2, 2009 at 9:34 am | Permalink

      Must mean it’s a very, very sinful dog.

  4. newenglandbob
    Posted October 2, 2009 at 9:15 am | Permalink

    The linking of disparate phenotype traits to a single gene or multiple gene sets appears to be a nightmare to deduce. Whether artificial or natural selection, a small change can have far reaching changes that will require thousands of studies to understand even one entire genome. I guess this means that the science of biology does not disappear any time soon due to lack of work.

  5. Chris
    Posted October 2, 2009 at 9:39 am | Permalink

    Totally off topic, but has anyone seen the “review” of Jerry’s book and Nick Lane’s ‘Life Ascending’ at
    It’s a very odd review….

    • hempenstein
      Posted October 2, 2009 at 3:51 pm | Permalink

      Indeed odd, to apparently come from the side of science yet refer at the end to evolution as the flickering light of a hand-held candle!

  6. Posted October 2, 2009 at 9:44 am | Permalink

    Please post something about cat coats.

  7. Posted October 2, 2009 at 10:06 am | Permalink

    Thanks so much for this post, I actually work on coat color and structure in other mammals and am currently carrying out a project on fgf5. What is even more interesting is that fgf5 is reported as responsible for the “long hair” phenotypes in many other mammals (cat, mouse, goat, sheep and more)

  8. Posted October 2, 2009 at 10:24 am | Permalink

    And that is a much more powerful explanation than anything the creationists can come up with.

    Faint praise, since “Mr. Magic Man did it” isn’t very powerful.

    Better is that it really does explain, in much the same way that forensics does.

    Glen Davidson

    • whyevolutionistrue
      Posted October 2, 2009 at 10:40 am | Permalink

      You’re absolutely right, Glen. In fact, I used precisely that analogy in a lecture to 600 First Year students this morning – if evolution were to be tested in a court of law, the DNA evidence alone would be enough to return a verdict of “proven”!


  9. Posted October 2, 2009 at 11:19 am | Permalink

    Are there no examples of phenotypes with a mutation only in keratin-71?

    Perhaps it would be short and curly (identical to F)?

    And how about coat color and patterning? Is variation in the genes coding for that also a recent occurrence? Any studies on that?

    • articulett
      Posted October 2, 2009 at 8:41 pm | Permalink

      I was wondering that too… there is no – – +

      I have a shepherd-looking dog with sort of medium fur… (and an all black tongue).

  10. Confused
    Posted October 2, 2009 at 11:43 am | Permalink

    Heh – I saw a talk by Elaine Ostrander, presenting this data (over the phone, since she couldn’t make it in person!) at the ISDB in Edinburgh last month. It was a fun talk.

  11. Posted October 2, 2009 at 11:51 am | Permalink

    Seems like a pretty ruff take on the subject.

    Sorry, couldn’t resist… Quite fascinating really.

  12. Ken Ham Fan
    Posted October 2, 2009 at 1:35 pm | Permalink

    This only shows MICRO evolution.
    All these dogs are from the same KIND.

    • newenglandbob
      Posted October 3, 2009 at 7:39 am | Permalink

      There is no such thing as KINDs. Ken Ham is an ignorant liar and his web site is nothing but ignorant garbage.

  13. Hempenstein
    Posted October 2, 2009 at 6:33 pm | Permalink

    Perfect that the current thread is about dogs since I need a segue to wolves. I hope everyone is or will have a chance to watch Ken Burns’ National Parks series. The last episode, that I’m watching now, has a great segment about Adolph Murie’s studies of the wolf while a Park Service agent, that led to protection of the species within the parks.

    The whole series has been full of examples of the results of efforts of single individuals, and has also been commendable in playing up the age of the earth here and there while avoiding any hint of creationary blather. An excellent production!

    • Hempenstein
      Posted October 2, 2009 at 6:34 pm | Permalink

      Errata: that should be Adolf Murie

  14. David Burbridge
    Posted October 3, 2009 at 5:23 am | Permalink

    Dog breeds are less than 200 years old?

    Yeah, right. The name ‘King Charles Spaniel’ ought to suggest a doubt or two about this. Or Shakespeare’s Macbeth:

    “Ay, in the catalogue ye go for men,
    as hounds and greyhounds, mongrels, spaniels, curs, shoughs, water-rugs, and demi-wolves are clipt all by the name of dogs.”

    Major breeds like greyhounds, wolfhounds, mastiffs, and spaniels all go back much further than 200 years, though of course the modern pedigree breeds are not identical to their ancestors.

    • whyevolutionistrue
      Posted October 4, 2009 at 7:46 am | Permalink

      See today’s post, David.

  15. Posted October 4, 2009 at 2:59 am | Permalink

    The list there doesn’t have KRT71 “by itself” (-,-,+).

    One wonders if it only occurs with at least one of the other two, or if it just didn’t contribute enough to the variation (perhaps because it’s rare).

2 Trackbacks/Pingbacks

  1. […] October 2, 2009 Jerry Coyne’s blog has a good post (written by guest blogger Matthew Cobb) summarizing a new Science article about the genetics of dog coats. [link] […]

  2. […] a previous post on the genetics of dog coat variation, I stated that ‘dog “breeds” are an incredibly recent invention – less than 200 […]

%d bloggers like this: