Last month a member of the Health and Genetics Committee called my attention to an organization called the Institute for Canine Biology (ICB), and in particular its 16-point manifesto entitled “Why You Need Population Genetics:  The “Elevator Pitch.”  During the discussion that followed, I expressed my opinion about the organization and its manifesto to the committee by email.  Since then, a couple of Deerhounders have asked me about the ICB, so I thought it would be a good idea to devote this month’s column to the organization.

by John Dillberger, DVM

Reprinted from the March/April 2015 Claymore

The ICB is a virtual organization, rather than a bricks-and-mortar one, so this website is the only way to visit them:  http://www.instituteofcaninebiology.org.  The ICB was founded in 2012 by Carol Beuchat, who earned a PhD in 1982 in physiology, with minors in ecology and physiological ecology.

The ICB defines itself as “an independent international consortium of canine biologists,” and the consortium includes some prestigious members from the USA and Europe.  The ICB’s mission seems to be captured in a program called “Breeding for the Future,” which was recently launched and which aims to “assist breeders in implementation of modern, scientific principles in their breeding programs by providing courses, written materials, computer simulations, and videos.”  In other words, the ICB is an education and service organization catering to purebred dog breeders.

On the ICB’s home page is this quote from Henry David Thoreau:
“To know that we know what we know, and that we do not know what we do not know, that is true knowledge.”

Perhaps this is meant to convey that the ICB intends to be balanced in discussing issues and appropriately humble about the advice it gives.

I found myself wondering why Ms. Beuchat created the ICB in the first place.  After reading through much of the ICB’s material, and particularly her blogs, I think I understand.  She believes that many (perhaps all) dog breeds have been inbred to the point of harming the breed itself and possibly threatening the breed’s very existence.  She believes that this inbreeding problem is reflected in declining health and reproductive fitness of purebred dogs in general and the large number of “genetic diseases” in various breeds.  She also believes that this inbreeding problem can be addressed only if breeders deliberately start doing the opposite—mating dogs that are as unrelated as possible.

There are two corollaries to this idea.  First, breeders will need to apply population genetic techniques to plan matings between dogs.  Access to these techniques is one of the main services that the ICB hopes to provide.  Second, and more importantly, breeders will have to stop thinking in terms of creating or maintaining a particular line within their breed and instead think of their breed as a single population.  Breeders will need to plan matings based on what is best for the breed as a whole, not on what is convenient, popular, sentimental, or likely to produce offspring that excel in the show ring or in a performance event.  This raises two fundamental questions:  what is best for a breed as a whole, and who decides?

As already noted, the ICB believes that what is best for a breed is for a mated pair of dogs to be as unrelated (and therefore hopefully as genetically different) as possible.  At the extreme, and for breeds with small populations in which all the dogs are to some extent related, this could involve opening studbooks and judiciously outcrossing with other breeds.  The ICB believes that this deliberate mixing of dogs within a breed—the opposite of the way lines and breeds themselves were created—is the only hope for improving breed health and saving rare breeds.  I am not ready to buy into the ICB’s premise yet, but I believe that I am stating it fairly.

The thornier real-world question is who decides what is best for a breed?  Ideally, it would be the breeders themselves, acting in unison.  But I think this is utopian.  Breeders would have to subordinate the idea of establishing a line of dogs or breeding for success in shows or performance events to the greater goal of genetic diversity.  It would require that breeders be forthright and honest about the good and bad phenotypic traits and genotypes of their dogs, and that such information be freely and conveniently available to other breeders.  (What I am describing is a health registry.)  It would require that breeders enlist the help of population geneticists like the ICB to help analyze potential matings.  If the ICB is right, then such a program undertaken for long enough, and with a dose of luck, would improve the overall health, fitness, and robustness of the breed.

As I said, I have some reservations about the ICB.  Those come partly from their manifesto, which you can find on their website and which I have reprinted at the end of this column.  As you can see it was written by Ms. Beuchat.  I suggest you pause now and read the elevator pitch in its entirety, then come back here and continue.

Calling the manifesto an elevator pitch makes clear that it is a sales pitch for the ICB’s ideas.  Any time one is presented with a sales pitch, it’s a good idea to listen or read carefully and critically, not only for what is said, but for what is not said.  This month I want to do just that with this elevator pitch, and dissect it point by point.  I do so in the hope that you, too, will carefully consider not only what the ICB proposes as solutions, but also what the organization sees as the problem.
Some of the ICB’s premises are ones with which everyone can agree.  Popular sires are used too often at some times in some breeds, and this creates an intentional genetic bottleneck or chokepoint.  Dog populations also experience unintentional bottlenecks, such as when their popularity wanes or when circumstances like a world war intervene.  Breeders do sometimes inbreed too often and too closely.  Inbreeding can increase the likelihood that a trait controlled by a recessive gene will begin to appear more often, be it a desirable or an undesirable trait.

Some of the ICB’s other premises are ones with which reasonable people may disagree.  For example, the ICB states that you cannot remove a single gene (allele) from a population and that in fact, you cannot even select for or against a single gene.  They also state that breeding for homozygosity of some traits ends up producing homozygosity of all traits, and that homozygosity is “the kiss of death for the immune system.”  And the ICB also believes that using DNA testing to remove disease genes from the population will not make dogs healthier.  I disagree with each of these ideas.

Before I begin going through the ICB’s elevator pitch point by point, I need to be clear on terminology, particularly the use of the words “gene” and “allele.”  An allele is a specific version of a given gene.  Some genes have only one allele.  Other genes have many alleles.  Not all breeds will have all possible alleles of a given gene.  For example, Deerhounds have a limited set of alleles for the gene(s) that govern coat color, so that no Deerhound has the brown, wheaten, or white coat found in other breeds.

So with that in mind, here is my point-by-point analysis of the ICB’s elevator pitch.

1)    All the useful genetic variation your breed will ever have was in the dogs that founded the breed. This genetic diversity is finite.
This is true.  But it is important to remember that this genetic variation/diversity consists of all of the alleles, not just all of the genes.  There are far more alleles than genes, as already explained.

2)    Each generation, alleles can be lost by chance (this is called “genetic drift”) and also through artificial selection by breeders, who select for dogs with the traits they like and remove other dogs from the breeding population.
Well, actually, one cannot lose an allele in a single generation (either by chance or selection) unless it is present only in dogs that are not bred.  Thus, the more dogs that are bred in each generation, the less likely that one will lose an allele from the breed, and conversely, the fewer dogs that are bred each generation, the more likely that one will lose an allele from the breed.  Also, the likelihood that an allele will be lost is inversely related to how widespread it is in the breed.  Rare alleles are easiest to lose.

3)    Because the stud book is closed, genes that are lost cannot be replaced.
This just re-states #1 in a different way.

4)    So, from the moment a breed is founded and the stud book is closed, loss of genetic diversity over time is inevitable and relentless.
This is the third statement of a syllogism comprised of points #1 and #3, and point #2, but it’s a little bit sloppy.  The syllogism should go like this.  If a breed can never gain new alleles (#1/#3) but can lose alleles in each generation (#2), then from the moment a breed is founded, it can lose genetic diversity.  This loss is not inevitable and relentless, but will instead be a function of how many dogs are bred each generation and the distribution of alleles in the members of each generation.  The fewer dogs that are bred, the greater the chance of losing an allele.  The scarcer the allele is in any generation, the greater the chance that it will be lost.

Of course, breeds can gain alleles by way of spontaneous mutation, but this is a relatively rare event and so it is OK to set it aside for this discussion.  And of course, another way that breeds can gain alleles is by outcrossing to a different breed.

5)    You cannot remove just a single gene from a population. You must remove an entire dog and all the genes it has.
Hold on a minute!  Are we talking about removing an allele (what Ms. Beuchat means by gene here) or an entire dog from a breeding population?  The second statement is true in a way, but what she means is that if you remove a dog from the breeding population, then you remove his contribution to that population—the alleles he might have passed on.  For common alleles, this doesn’t matter, since removing one dog might mean that there are 198 copies of the allele in the breeding population instead of 200 copies.  The rarer an allele, the more it matters whether or not one of the dogs carrying it is not used for breeding.
If the allele is a “bad” one, such as an allele that results in a health problem, then this is a good thing.  If just a few dogs carry that allele, and by chance or intent, none are used for breeding, then you will eliminate the allele from the breeding population and therefore from the breed.  Thus, you can remove just a single allele from a population.  Heck, that is how breeds were created in the first place—by removing lots of alleles from a population!

6)    You cannot select for or against a single gene, because genes tend to move in groups with other genes (this is called “linkage”). If you select for (or against) one, you select for (or against) them all.
This is ridiculous!  Of course you can select for or against a single allele (what Ms. Beuchat means by gene here).  She is just pointing out that in doing so, you also may be selecting to a lesser extent for or against other alleles.  The closer two genes lie together, the more tightly linked their alleles will be in a given dog, and the more likely it is you will select against both alleles if you select against one.  But linkage is not 100%.  And more importantly, linkage between alleles in a given dog is not the same thing as the linkage between genes in the population.

Let me illustrate with an example.  Let’s say that one gene has two alleles (A and B) and the gene right next to it has six alleles (U, V, W, X, Y, and Z).  Assuming the alleles are randomly distributed in the population (they won’t be, but it’s OK for our example), then a given dog with allele A might have one of six genotypes:  AU, AV, AW, AX, AY, or AZ.  But in fact, a dog has two copies of each gene, and so our hypothetical dog could have one of 21 genotypes:  AU/AU, AU/AV, AU/AW, AU/AX, AU/AY, AU/AZ, AV/AV, AV/AW, AV/AX, AV/AY, AV/AZ, AW/AW, AW/AX, AW/AY, AW/AZ, AX/AX, AX/AY, AX/AZ, AY/AY, AY/AZ, or AZ/AZ.

Now imagine you decide to select against allele A.  Which allele of the next-door gene are you also selecting against?  It could be any of them!  Given the law of averages, you are selecting equally against each of the six alleles of the other gene, and the impact on any one allele will be relatively small.  The farther away a gene is from the one for which you are intentionally selecting against a given allele, the less the impact.

The same is true, of course, if you are selecting for a given allele of a particular gene.  All breeders know this.  That is why it is so hard to get a bunch of good traits (alleles) in a single dog.  Even if the governing genes are closely linked, they will carry the desired alleles in only a handful of offspring, and it might take many generations to get the desired alleles into a single dog.  And once you do, you’re still not done.  For the dog with the desired set of alleles to reliably pass them on to his pups, the alleles must be on both chromosomes—that is, the dog must be homozygous for each of the desired alleles.

In any event, the point is that selecting for or against one allele of a given gene is not only possible but is the goal of all breeders, who are trying to do this with hundreds of genes at each breeding.  By doing so, they are not automatically also selecting for or against certain alleles of other genes.

7)    Breeding for homozygosity of some traits breeds for homozygosity of all traits.  Homozygosity is the kiss of death to the immune system.  And by the way, as genetic variability decreases, so does the ability of the breeder to improve a breed through selection, because selection requires variability.
This stopped me in my tracks.  Ms. Beuchat has gone off the deep end here.  Homozygosity means that the same allele is present in both copies of the gene that a dog has.  From the previous example, you can see that breeding for homozygosity in one gene (AA) does not mean you will automatically get homozygosity in the gene next door (for example, UU).  If that were true, then breeding would be easy.  But breeding is not easy at all.  If AA and UU are what you want—the best allele of each gene—then it might take your entire lifetime to finally get a dog that is AA and UU.  You certainly won’t end up with UU automatically simply by selecting for AA, unless the only allele of the second gene in the breeding population is U.  And if that is so, then all your dogs already are UU.

Moreover, once you get the dog that is AA and UU (AU/AU), he is only half the story.  Unless his mate also is AU/AU, then the pups will be a mish-mash.
As for homozygosity being the kiss of death for the immune system—hogwash!  Some alleles will “improve” the immune system, and other alleles will “impair” the immune system.  A dog that is homozygous for alleles that improve his immune system has not gotten the kiss of death.  Instead, he has gotten the kiss of life.

Ms. Beuchat’s last statement also is not true, and the way in which it is false is extremely important.  The statement should read like this:  As genetic variability decreases, so does the ability of the breeder to improve or harm a breed through selection.  If an allele has been lost to the breed, then it is no longer available for use.  If the lost allele resulted in a beneficial trait, then the ability to improve the breed by having that beneficial trait is gone.  But if the allele resulted in a harmful trait, then the ability to harm the breed by having that trait is just as surely gone.  This is the whole idea behind trying to eliminate alleles that result in (or increase the risk of) a harmful trait.  Nobody wants to lose an allele that controls a desirable trait.  Everybody wants to lose an allele that controls an undesirable trait.  This is just breeding.

8)    The consequences of inbreeding (in all animals) are insidious but obvious if you look – decreased fertility, difficulty whelping, smaller litters, higher puppy mortality, puppies that don’t thrive, shorter lifespan, etc. Genetically healthy dogs should get pregnant if mated. They should have large litters of robust puppies, with low pup mortality. Animals that cannot produce viable offspring are removed by natural selection.
Last time I looked, lots of breeds (of dogs and every other species) were getting pregnant if mated and producing large litters of robust offspring with good survival.  Some breeds of animals have been doing so for hundreds, perhaps thousands, of generations.  The trick is to have gotten a good set of alleles for the genes that affect fertility, pregnancy, birthing, development, etc. in the first place, or to have obtained a good set of such alleles over time.  How?  By selecting for them!

9)    Mutations of dominant genes are removed from the population if they reduce fitness. Mutations of recessive alleles have no effect unless they are homozygous. So rare alleles are not removed, they are inherited from one generation to the next, and every animal has them.  Lots of ’em.
Oh how I wish that dominant alleles were removed from the population if they reduced fitness!  But alas, the only filter is reproductive fitness.  If a dominant allele reduces reproductive fitness to zero, then the allele will be eliminated right away.  If it reduces reproductive fitness even a little bit, it will gradually fade away until it disappears.  But if a dominant allele reduces fitness in any other way, it will not be selected against.  Such an allele might increase the risk of osteosarcoma, or bloat, or cardiomyopathy, or neck pain, or pneumonia, or anything else that doesn’t affect reproductive fitness because the effects of the allele manifest after a dog already has reproduced.

That last bit about rare alleles not being removed is just plain wrong.  Remember that Ms. Beuchat’s point #2 was that “each generation, alleles can be lost by chance [or] through artificial selection by breeders.”  And as I noted, the rarer an allele is, the more likely it is to be lost.  I think what Ms. Beuchat was trying to say is that recessive alleles are not removed.  But of course, that is wrong, too.  Alleles can be lost by chance or intent regardless of whether they are dominant or recessive.

10)    If you create a bunch of puppies from your favorite sire, you are making dozens of copies of all of the bad alleles in that dog (which were never a problem before because they were recessive; see 9) and spewing them out into the population.  Now, a (previously) rare mutation will become common, its frequency in the population increases, and the chances go up that some puppy will be produced that is homozygous (has two copies of that bad allele) – and homozygous recessive alleles are no longer silent.
Oh, where do I begin?!  Perhaps the easiest way is to change the word bad to good.  Ms. Beuchat could as easily have written this:

If you create a bunch of puppies from your favorite sire, you are making dozens of copies of all of the good alleles in that dog (which were never recognized before because they were recessive) and spreading them out into the population.  Now, a (previously) rare mutation will become common, its frequency in the population increases, and the chances go up that some puppy will be produced that is homozygous (has two copies of that good allele) – and homozygous recessive alleles are no longer silent.

Somehow it doesn’t sound so bad when put like this.

Point #10 is a hypothetical situation.  Whenever presented with one of those, you need to work through it.  Ms. Beuchat starts with a hypothetical sire that has single copies of many bad alleles (what she calls “mutations”), which are rare in the population.  I know he has only single copies because she says they were “never recognized before,” and that can happen only if the allele is both recessive and present in only one copy.  I know they are rare in the population because she says so.  To take the most extreme example, let’s assume that each bad allele is present only in this sire.  Since the sire has only one copy of each bad allele, his pups will have a 50:50 chance of getting each one, on average.  Some may get all the bad alleles.  Some may get none of them.  Most will get some subset of the bad alleles.

Once the pups grow up, then it seems obvious that each bad allele will be more common in the breeding population than it was before.  But is that true?  Again, let’s take the most extreme example and assume the dog produces 50 pups, and they are all male.  Now there are 50 potential sires, where before there was only one.  And on average, each bad allele is present in 25 of those potential sires.  That is a lot more copies of each bad allele in the population—but there are also a lot more potential sires in the population.  In the first generation—the original sire—each bad allele had a frequency of 100%.  In the second generation of sires, the frequency of each bad allele is only 50%.  Each bad allele has become less common, not more common!

And if we choose one of those 50 sires and go through the exercise for another generation, the frequency of each bad allele that he carries (half of the ones his sire carried) gets cut in half again!  In other words, the bad alleles become less common, not more common.  Their frequency in the population decreases, rather than increases.  Again, all breeders know this, if they simply substitute the term “good allele” for “bad allele.”  Sometimes one wants desperately to hold onto a good set of alleles in a sire, only to have them get more and more diluted with each breeding.

As this dilution continues with each successive generation, and the frequency of a bad allele declines in the population, the chances become greater and greater that the bad allele may be lost entirely from the population (see my response to #2).  To think of it another way, if one picked a sire at random to use for each successive breeding, there would be a 100% chance he had the bad allele in the first generation, 50% chance he had it in the second generation, 25% chance he had it in the third generation, etc.

11)    So, genetic disorders caused by recessive alleles don’t “suddenly appear” in a breed.  The defective gene was probably there all along.  Make a zillion copies, and suddenly you have a disease.
Of course the recessive allele was there all along.  But it is not true that “making a zillion copies” will lead to its becoming homozygous in more dogs and showing up as a disease.  As I explained in my response to #10, one cannot make a zillion copies of the recessive allele without making several zillion dogs, so that the frequency of the allele in the population actually goes down, rather than up, with each successive generation.

The frequency of a recessive allele can go up over time, if the dogs used for breeding for several successive generations are carriers.  As I have explained, the odds of that happening by chance go down with each generation.  But of course, breeding is done by intent, not by chance.  A popular sire may be used for dozens of breedings in a given generation.  Indeed, thanks to frozen semen and artificial insemination, a popular sire may be used for dozens of breedings over several generations.  It is this practice, not some inevitable law of nature, which can cause a recessive allele to increase in frequency.  But then, that is the point of using a popular sire—to increase the frequency of his good alleles in the population.
One way to avoid increasing the frequency of an undesirable allele in a population is not to use any sire too often, something the ICB advises and something with which I agree.  But this can be difficult in breeds with small populations, like Deerhounds, where there are not only relatively few potential sires but even fewer that complement any particular bitch.  And of course, breeders are always trying to “breed the best to the best,” which tends to pull them toward the same few sires.

Another way to avoid increasing the frequency of a bad allele in a population is to be able to identify dogs that carry it and exclude any of their offspring that also carry the allele from the next generation’s breeding population.  That way, one could use a popular sire without fear of sowing the undesirable allele into future generations.  This is the idea behind genetic tests for undesirable alleles—to use the test to reduce the frequency of bad alleles without having to also reduce the frequency of good alleles.  Ms. Beuchat does not see things this way…

12)    Using DNA testing to try to remove disease genes from the breed will not make dogs healthier (see 2, 5, and 6).
I cannot help but get angry at this statement.  The harm it could cause is hard to calculate.  Even on its face, the statement is absurd.  How could removing an allele that causes (or increases the risk of) a disease in a breed NOT improve the health of the breed?

I think what Ms. Beuchat means to say is that removing a disease-causing allele will not make a breed healthier because, in removing that bad allele, one will also be removing other desirable alleles, and so one will make dogs healthier in one way but unhealthier in another.  But as I explained in my responses to #2, #5, and #6, that is not true.

13)    The breed will continue to lose genes every generation (by chance or selection) until the gene pool no longer has the genes necessary to build a healthy dog.
This is a pretty dire prediction, meant to create sense of urgency and doom.  There is truth to what Ms. Beuchat says, in that alleles will be lost over time.  But the trick is to try to minimize the loss of good alleles and maximize the loss of bad alleles.  To the extent that one succeeds, one actually improves the gene pool, even though it might contain fewer total alleles.

And what is the best tool for speeding the loss of bad alleles?  DNA testing!  A breeder cannot avoid what s/he cannot see.  DNA testing makes the invisible visible.  Sure, this tool can be mis-used, just like any other.  But just because some people hurt themselves with a chainsaw does not mean we should not use chainsaws to rescue people trapped in collapsed buildings.

14)    At this point, the breed might look wonderful (because of selection for type), but it will suffer from the ill effects of genetic impoverishment – inbreeding depression, diseases caused by recessive alleles, increased risk for cancer, etc.
I think Ms. Beuchat meant this to be part of #13, so that the phrase “at this point” means the point at which a breed no longer has enough of the right alleles left “to build a healthy dog.”  But point #14 is just ridiculous.  I do not know of anything to suggest that any health problem is cause by “genetic impoverishment,” whatever that means.  It sounds like a term coined for a sound bite on the nightly news.

The best thing that could happen for a breed suffering from a “disease caused by a recessive allele” would be for someone to develop a genetic test for the allele so that it could be eliminated from the population.  This would, by definition, further impoverish the breed genetically, since there would be one fewer allele in the population.  But it would most certainly improve the breed’s health.

15)    The health of individual dogs cannot be improved without improving the genetic health of the breed.  The only way to improve the genetic health of the breed is to manage the health of the breed’s gene pool.
This is catchy and sounds like it ought to be true.  But is it?  Aren’t some subpopulations (lines or regional populations) within a breed more or less healthy than others?  I agree that managing the health of an entire breed’s gene pool is one way to improve the health of individual dogs, but I do not believe it is the only way.

16)    Population genetics provides tools for the genetic management of breeds or other groups of animals.  Breeders CAN improve the health of the dogs they breed if they understand and use them.
This is true.  Breeders of livestock and laboratory animals have been using these tools for a long time.  They work.  But it requires that one manage the breeding of a large population.  In the case of Deerhounds, it would require that someone manage the breeding of, say, all North American Deerhounds or (ideally) all Deerhounds in the world.  No individual breeder can do this.

As I said at the beginning of this column, I believe this is what the ICB has in mind—to offer their services to help manage the breeding of entire breeds.  If provided with pedigree and health information on individual Deerhounds, they could use population genetic analyses to advise individual breeders or, if there were a global Deerhound organization, to advise that organization.  Of course, the unspoken understanding is that breeders or the global organization would be more interested in the long-term health of Deerhounds as a breed than in anything else.  There’s a bit of the utopian mindset here, but more power to them.  They are at least pointing out that too much inbreeding can get a breed into trouble, and offering to help breeders and breed clubs who would like to reduce inbreeding.
_______________________

I believe that the ICB has the best interests of purebred dogs and breeders at heart.  But I think the organization tends to present a problem (too much inbreeding in some breeds at some times) as if it were a disaster and an emergency.  Frightening people is not a way to help them make thoughtful decisions.  I also do not agree with some of the ICB’s underlying premises, particularly that the health of purebred dogs is bad.  I think that idea is founded on a misperception.  In a blog entitled “Why do dogs have so many genetic disorders?” Ms. Beuchat shows a graph to illustrate that genetic diseases are more common in dogs than in 12 other species.  As she says:

“Dog outstrip all other species in the number of documented genetic disorders, with 619, followed (far behind) by the domestic cow with 443.  From there the numbers decline (219 in the horse, 18 in the guinea pig, 7 in the ferret).  A number of these animals are commercially bred (e.g., cows, sheep, pigs), while others are kept as pets (e.g., cat, guinea pig), or for food or wool (e.g., goat, llama, alpaca).”

But I think Ms. Beuchat has overlooked the more obvious explanation for these data.  The reason that more genetic disorders have been identified in dogs than the other species on her list is because more research effort has gone into looking for those disorders.  You can confirm this by looking at two species that she doesn’t list that also have been the focus of much research into genetic factors that influence disease:  mice and humans.

If the large number of know genetic disorders in dogs were due to a century of inbreeding, then one would expect many fewer genetic disorders in laboratory mice (where breeding colonies have long been managed using population genetics) or humans (where breeding is random).  But in fact, the opposite is true.  In mice, there are close to 1000 known genetic disorders.  More than 700 of these are maintained as separate mouse strains in the Mutant Mouse Colony at The Jackson Laboratory in Maine (http://mousemutant.jax.org/).  In humans, more than 2000 genetic disorders had been identified by the year 2000, and the number has climbed steadily since then.  You can find the current list at a website called the Online Mendelian Inheritance in Man (OMIM) at http://www.ncbi.nlm.nih.gov/omim.  The situation in humans is particularly relevant, as it suggests that even when inbreeding is not practiced, genetic disorders are abundant.

As I have tried to illustrate in this column, when presented with a “call to action” such as the ICB’s elevator pitch manifesto, you need to read carefully and critically and then apply a strong dose of common sense.  Don’t take statements at face value; instead, consider the author’s underlying assumptions and motivations.  Recognize that even if a blog appears on a sophisticated website maintained by an official-looking organization with prestigious members, that it is still an opinion piece.  And finally, remember that blog writers make mistakes.

Why You Need Population Genetics: the “Elevator Pitch” 
by Carol Beuchat, PhD

• All the useful genetic variation your breed will ever have was in the dogs that founded the breed. This genetic diversity is finite.
• Each generation, alleles can be lost by chance (this is called “genetic drift”) and also through artificial selection by breeders, who select for dogs with the traits they like and remove other dogs from the breeding population.
•  Because the stud book is closed, genes that are lost cannot be replaced.
• So, from the moment a breed is founded and the stud book is closed, loss of genetic diversity over time is inevitable and relentless.
• You cannot remove just a single gene from a population. You must remove an entire dog and all the genes it has.
• You cannot select for or against a single gene, because genes tend to move in groups with other genes (this is called “linkage”). If you select for (or against) one, you select for (or against) them all.
•  Breeding for homozygosity of some traits breeds for homozygosity of all traits. Homozygosity is the kiss of death to the immune system. And by the way, as genetic variability decreases, so does the ability of the breeder to improve a breed through selection, because selection requires variability.
• The consequences of inbreeding (in all animals) are insidious but obvious if you look – decreased fertility, difficulty whelping, smaller litters, higher puppy mortality, puppies that don’t thrive, shorter lifespan, etc. Genetically healthy dogs should get pregnant if mated. They should have large litters of robust puppies, with low pup mortality. Animals that cannot produce viable offspring are removed by natural selection.
• Mutations of dominant genes are removed from the population if they reduce fitness. Mutations of recessive alleles have no effect unless they are homozygous. So rare alleles are not removed, they are inherited from one generation to the next, and every animal has them.  Lots of ’em.
•  If you create a bunch of puppies from your favorite sire, you are making dozens of copies of all of the bad alleles in that dog (which were never a problem before because they were recessive; see 9) and spewing them out into the population.  Now, a (previously) rare mutation will become common, its frequency in the population increases, and the chances go up that some puppy will be produced that is homozygous (has two copies of that bad allele) – and homozygous recessive alleles are no longer silent.
•  So, genetic disorders caused by recessive alleles don’t “suddenly appear” in a breed. The defective gene was probably there all along. Make a zillion copies, and suddenly you have a disease.
• Using DNA testing to try to remove disease genes from the breed will not make dogs healthier (see 2, 5, and 6).
• The breed will continue to lose genes every generation (by chance or selection) until the gene pool no longer has the genes necessary to build a healthy dog.
• At this point, the breed might look wonderful (because of selection for type), but it will suffer from the ill effects of genetic impoverishment – inbreeding depression, diseases caused by recessive alleles, increased risk for cancer, etc.
• The health of individual dogs cannot be improved without improving the genetic health of the breed.  The only way to improve the genetic health of the breed is to manage the health of the breed’s gene pool.

Population genetics provides tools for the genetic management of breeds or other groups of animals.  Breeders CAN improve the health of the dogs they breed if they understand and use them.

Copyright © 2013 Carol Beuchat


This document may be reproduced without permission if accompanied by the copyright information.