Genetics
My dog is a quarter wolf
No, no it’s not – or at least, you can’t know whether this is true.
Actually, this isn’t about my dog at all. I don’t have a dog with a wolf grandparent. This blog is about the idea that we can say any particular dog is “87.5% wolf” or “one-quarter Jack Russell” or most of the other fractional values we assign to describe a dog’s heredity. This concept is confusing and hard to explain, so really, I’m writing this as a blog so I have a handy reference for when it comes up in discussion.
Some very basic genetic principles applying to sexual reproduction:
1. When two parents have babies, each baby’s DNA is half mom’s and half dad’s.
2. We have chromosomes, which are conglomerated chunks of DNA, broken up into segments called genes. (There is also “filler” which does its own complicated stuff.) Chromosomes come in two matched sets, with congruent addresses on the matching chromosomes.
For example, on dogs, gene locus R306ter contains genes that code for the presence or absence of “e-locus yellow” — the color of Golden Retrievers and yellow Labradors, among others. The dominant gene, E, is a slightly different shape than the recessive gene, e. If the dog has E on both chromosomes, its coat color will be the underlying black or chocolate. If the dog has E on just one of the chromosomes, the same result occurs, because E is dominant (will express and the recessive version will not express). If BOTH chromosomes have the recessive e, then this protein will be expressed, creating that yellow-gold color of some of our favorite breeds. All Goldens are e/e, so they will never show up in chocolate or black (but you can tell what’s there by looking at their nose leather). Labs can have any combination, so you can get yellow or chocolate or black.
3. One of those gene versions came from Mom. The other one came from your Dad. This is true for every single gene locus, whatever the gene does.
4. If we are talking about two individuals able to reproduce together, we know that they have an awful lot of DNA in common with each other. This means that for a whole lot of gene loci, both Mom and Dad are contributing the same version of the gene — that is, two protein-generating chunks of DNA which create the same protein. Within a breed, there is even less variation from one individual to the next. But still, some of the gene loci could have different versions of the gene. For example, Mama Labrador might have E/e at her R306ter locus, while Daddy Lab might have e/e at his R306ter locus.
5. Back to the puppies of these two Labs: here is where those Punnett squares from high school come in. Blue collar puppy got E from Mama Lab (could have gone either way) and e from Daddy Lab (only option available). She is E/e, like Mama. Since the dominant version of the gene, E, is present, the recessive e/e yellow color cannot express and Blue collar girl will be either chocolate or black. (That’s another whole puzzle which we are ignoring for now.) Now let’s look at Red collar boy. He got e from Mama (could have gone either way), and e from Daddy. He is therefore e/e and will express a yellow or golden coat color. He will be a yellow Lab.
Now, let’s just imagine that a bad thing happens and Blue collar and Red collar have an unchaperoned fling and an accidental litter. With respect to yellow coat color, we have one E/e parent and one e/e parent… just like these two dogs’ own parents. And likewise, these new puppies, Purple and Green, could end up either E/e or e/e. We would just say they are black (or chocolate) or yellow. We would not say that Purple (our third generation E/e) is “one quarter yellow.” Purple’s yellow coloring genetics are exactly the same as her Mama’s and her Mama’s Mama’s. She has a fifty percent chance of passing on E to one of her own puppies, and a fifty percent chance of passing on an e. Same as the others. And even more important, we would not say that Green (e/e) is “one quarter black” or even “half-black,” even though one of his parents and one of his grandparents are black. Because even though those dogs were black, we know that Green has ONLY the e version of the gene and that is why he is yellow colored. The fact that an ancestor had the other version which allowed black to be expressed has zero effect.
Now that we have worked through that thought experiment, switch the two original parents to Border Collie and Whippet. (This is a popular purpose bred sports mix in flyable, not to mention an old-fashioned lurcher.) Our first litter of puppies will have half BC genes and half W genes. Because this is a primary cross between two breeds, it is referred to as an F1 generation or F1 litter. So at every single gene locus, we have BC/W — that is, the Border Collie version of the gene, and the Whippet version of the gene. (See additional comment below.) These pups will tend to look a lot like each other because they are in fact half-and-half. Further, Whippets and BCs share plenty of DNA and have very few super-divergent traits anyway, so there won’t be genes creating traits that obviously belong to another breed to create different looking puppies.
Now, let’s imagine a breeding between two F1 “Borderwhippets.” (Related or not, doesn’t matter.) At every gene locus, each baby gets one gene from each parent. From each parent, Orange collar boy could get the BC or the W. So the possible combinations are:
BC/BC, or BC (mom)/W(dad), or BC(dad)/W(mom) or W/W. Or to put it another way, there’s a 25% chance that the combination at our sample gene locus will be all Border Collie. And there is a 25% chance that it will be all Whippet. And there is a one-half chance it will be half-and-half.
This is true of every single gene locus on Orange collar boy’s chromosomes.
If we assume that assortment is truly random (it could go like the flip of a coin for each and every gene locus), then we can estimate that Orange’s gene loci will be about 25% all Border Collie (the BC versions of the genes), 25% Whippet (the Whippet versions of the genes) and 50% mixed. If we take this average, it will generally average out, and these F2 puppies will sort of look like their parents. But, it’s possible for one of these F2 pups to get way more than half BC DNA, just because it happened to get the BC version of the gene from both sides. Or it could be way more Whippet. It’s even theoretically possible for an F2 puppy to get 100% Whippet genes and be genetically indistinguishable from a purebred Whippet.
This is astronomically unlikely, but it’s possible, just like it’s very unlikely you will flip heads 100 times in a row, but possible. And similar mixing or weird assortments can happen in F3, in F4 and so on.
Likewise, if you breed an F2 dog (half W genes, half BC genes) back to a purebred Whippet dad (W/W), your pups (generation F2a) might come out W(dad)/W(mom); or W(dad)/BC(m0m). In other words, we could see a purebred Whippet or another 50/50 Borderwhippet if the most unlikely case comes to pass. But most likely it will be somewhere in between, with the biggest probability at around 25% BC genes and 75% Whippet genes. So when we say a dog is “one quarter BC” because it had one BC grandparent, what we are really saying is that there is some statistical likelihood that the dog’s genes are one quarter BC. But reality could very easily be 24 or 21%, or 29 or 33%, and there is no way to know (at this time).
Here are some conclusions we can draw from this:
1. F2 and F3, etc. puppies cannot be said to be “one-quarter” anything, genetically. This is not paint mixing where every gene that has ever been present in any ancestor is in the pot. Instead, in every single generation, half of the candidate genes for this new pup’s genome don’t make the cut. (If it were paint, we could not go dive around in the pot and dig out every red molecule to get a purer green. All we can do is add more and more green so that there is too little red to see.) Once a particular version of a gene fails to make the cut, it can never crop up again. That’s why all Vizlas, Dogues de Bordeaux and Pharaoh Hounds are genetically chocolate (liver nosed) — the version of that particular gene which codes for black hair has been totally eliminated. Anyway, any claim that a dog is “one quarter” anything, or “92% wolf” is based on the paint-mixing model — which is not how genetic assortment works. You can be sure anyone who makes such claims simply does not understand genetics very well (and might avoid taking their other genetic claims too seriously).
As an aside, I have a hypothesis that a lot of the primarily flyball-bred sport mixes bred as height dogs from combinations of Border Collie and various terriers (Border, Jack Russell and Staffy Bull) are, in the later generations, much more “terrier” than their imagined “percentages” would suggest. I think this is because the terrier genetics contribute the small size and breeders continue to select for that small size, thus effectively preserving terrier-specific genes. For example, scroll down here to Blue Cedar’s Quasar, a very sweet little dog who is “said to be” half Border Collie based on his pedigree: http://www.portlandtailblazersflyball.club/#!our-dogs/gdy0f
As you can see, he looks exactly like a mixed up terrier, and is the size of one (under 15″ at the shoulder). Aside from being quite biddable and peaceful with other dogs, he also acts like one (he has many squirrel kills to his name, and his terrier scream is unmistakable.) He’s about an F4 mix. I have no doubt that if we could magically map every gene on his genome, he would favor terrier much more than 50%.
2. F2 puppies tend to be pretty consistent with each other and predictable. F3 and F4 puppies tend to be a lot more diverse and can be pretty bewildering if the original breeds that went into them are divergent (e.g. very different in build, leg length, or head shape).
I know of an F2 Borderwhippet litter of 8 puppies who all debuted in flyball around the same age, and all put in times between 3.4 and 3.6 seconds at their first tournaments. These are literally world-record class times which only a handful of dogs is capable of achieving to begin with. Getting an entire litter which could do this is a level of consistency most dedicated purebred breeders only dream of. Of course this is partly luck.
I realize that this is a complicated subject and many may not have made it this far. I suppose if I understood this better myself I could produce a shorter, more succinct explanation! I also realize that I have oversimplified a great deal and made some broad generalizations which will enrage people with more genetic background. For example, we did not talk at all about sex-linked traits (where there may not exist two copies of a gene because the X and Y chromosomes are not congruent shapes, among other things) or maternal DNA or the fact that actually, the vast majority of the DNA in any two dogs of any breed is virtually identical, and the number of genes that create breed differences are actually very few. It would be more accurate to say that in an F2 dog, half of the genes which may differ between breeds will be from one of the breeds, and half of the genes which may differ between breeds will be from the other, and these comprise maybe just a fraction of a percent of all the genes on all the chromosomes. But that is very cumbersome to say and does not add to the basic point.
I will be open to revising this if I have the energy and also if people have corrections (if I have made material mistakes which are not just oversimplifications). I am not a geneticist, just an interested amateur. Thanks for reading.