Let's say you're a responsible, preservation breeder (of course!), and you pride yourself in prioritizing health as a breeding goal. You carefully research pedigrees and ancestry before making mating decisions. You do all the relevant health testing (i.e., mutation testing and phenotype evaluation) for your breed. Yep, this sounds like what we want to see in a purebred dog breeder!

I'm now going to rattle your cage a bit.

What if there are things happening right under your nose that are causing significant deterioration of your breed's gene pool?  You realize that this would not be a good thing, and I'm sure you would want to know about it. Reducing the quality of your gene pool will affect breeding in every generation to follow. You rightly think this sounds like something breeders should be monitoring and mitigating if the goal is breed preservation.

One last point before you dive into the fun stuff. These mystery effects could be countering all the efforts (and time and money) you are investing in your efforts to breed better dogs. So not only is there possible damage to your gene pool, there is also a chance that your breeding program is circling the drain and you don't even know it.

What follows is one of the lessons (Module 13, unit 1) from of an online course in population genetics that I created specifically for dog breeders (ICB Managing Genetics for the Future). This unit comes rather late in the course (Module 13 of 15), so you will be missing the extensive background concepts and information leading up to this that would make it more meaningful for you. But I think you can get a lot out of it if you take the time to follow the instructions carefully as you work your way through. Think about your own breed as you do this, and I think you will see some fog clearing. Even better, throw in some numbers for your breed and see if your breed population is as healthy as you think it is.

​Above, all, have fun!

Genetic management: basic strategyThe reason we have domestic animals of particular breeds is because they are useful to us. Dogs have been put to work guarding, herding, hunting, hauling, fighting, controlling vermin, alerting, guiding, and more, and of course they have been companions as well. Breeding them deliberately allows us to produce that animals that best suit our needs through genetic selection for particular trait. Developing a breed with certain traits is essentially the process of creating a gene pool with the genetic material necessary for those traits as well as the variation necessary to improve or modify these traits as needed. Of course, if you can't breed domestic animals sustainably, your investment in their development and benefit from their use will be lost.

The development and sustainable breeding of dogs went along fine for thousands of years. Breeds that were needed were created, and those that were no longer needed went extinct. Genetic management was accomplished by doing what worked, selecting the best dogs for their purpose, culling the unfit, and finding dogs to breed to that would improve or add desired traits. Purpose and function drove selection, and the requirements that working dogs be both fit and inexpensive to keep insured that health was maintained. In the late 1800s, the purpose and means of breeding dogs significantly changed. Events originally intended for the evaluation of breeding stock turned into contests to identify the "best" specimens, then closed studbooks and "purebred" dogs were created.

This fundamentally changed the way dogs could be bred, and most significantly it removed the possibility of introducing new genes into the gene pool of a breed. The consequence has been the steady and dramatic reduction in both the genetic and physical health of these purebred dogs until we get to the situation we have today, in which the major topics of discussion among breeders are not the traits and quality of their dogs, but how to accomplish a successful breeding in the face of reduced fertility, smaller litters, high puppy mortality, and the increasing number of heritable disorders.

The health problems we have today in dogs are a direct and predictable consequence of a failure to properly manage the gene pools of the breeds. The truth of the biology is that closed gene pools lose diversity inexorably over time, and as homozygosity and homogeneity increase, health and reproductive success deteriorate. Eventually, inevitably, closed populations go extinct.

Sooner or later, if purebred dogs as we know them are to survive, the stud books must be opened and biologically appropriate methods of breeding must be allowed. Until then, breeders need to improve and protect the quality of their breed's gene pool, and in fact they will still have to do this after stud books are open.

The science of breeding and managing populations of domestic animals has developed considerably over the last 100 years, and especially in the last few decades. Genetic management strategies have been extensively explored and perfected in many kinds of animals in thousands of breeding experiments, and in fact some groups such as service dog organizations have adopted modern breeding methods that dramatically improve their ability to manage inbreeding while also producing dogs of reliable type and temperament. Breeders of all purebred dogs need to learn to do this too.

Wild animal species are closed populations too
The African elephant, Grey wolf, song sparrow, brook trout, and thousands of other familiar species of animals have been around for thousands of generations without becoming more and more inbred over time. In fact, most healthy wild animal populations have levels of inbreeding of only a few percent. How do they manage this?

Animals in a single population become inbred on each other. If you split a population, the two populations will also become inbred, but not in the same exact way. Different alleles will be lost by genetic drift, selection will be slightly different, and over several generations they will drift apart genetically. If an animal then moves from one population to the other, it can reintroduce some of the alleles that might have been lost. The ability of animals to migrate among isolated populations is how species minimize the level of inbreeding and depletion of the gene pool over time and, for the same reason, isolation of a population of animals so that migration cannot take place leads to increased inbreeding and genetic drift.


This can work the same way in domestic animals as it does in wild ones, and maintaining multiple genetic groups or lines within a breed is a useful management strategy if used properly. What this effectively does is create multiple open populations within a larger closed one.

The tradeoff is that smaller populations accumulate inbreeding more quickly and the rate of genetic drift is higher, so these things need to be taken into account when planning a breeding strategy.

Since breeding dogs aren't kept in herds or separate physical groups, these subpopulations exist on paper, and management is the process of monitoring the inbreeding and outcrossing as appropriate.

Simulating the genetics of multiple populations
We're going to use a computer simulation that will allow you to see how changes in population size, number of subpopulations, and migration rate (which for dogs is equivalent to a cross to another line) affect the genetics of the overall population.

Go to the PopGen Simulator-
http://www.radford.edu/~rsheehy/Gen_flash/popgen/

You will see two graph axes and a set of parameters you can fiddle with across the top.
​
Use these settings to start:
Pop. size = 50
A1 allele frequency = 0.5
# of Populations = 1
Number of generations = 200
and leave the "Fitness" boxes all at one, migration unchecked, mutation rates both as zero, and nothing for bottleneck.

Click "Go". In the top graph, you will see that you have lines for the frequencies of two alleles, A1 and A2. Since you set A1 to 0.5, A2 will be the same at the start (because A1 + A2 must equal 1).

In the graph below, you will see lines for how the genotypes change over time. Of course, this is from the Hardy-Weinberg equation that we talked about earlier.

Leave the settings and click "Reset" then "Go" several times, and you will see that you get a different outcome every time. This is because the simulation is modeling the random inheritance of A1 vs A2 each generation. You will see that one allele or the other can disappear entirely from the population, which means the other one is "fixed" - i.e., every individual is homozygous for the remaining one.

Now, let's do some experiments and collect some data.

1) Run this simulation 10 times with the same settings and write down which allele becomes fixed and the generation (roughly) when that happens.

Change the population size to 200, and run 10 times recording the data as above.

Change the population size to 25, and run 10 times recording the data as above.

From you data, you can see that a small population is genetically unstable, and the smaller it is the more unstable and less predictable it is.

2) Change population size to 200, and change number of populations to 2. Repeat as above.
You'll see that the display changes when you do this. Now, we don't get the genotype graph, but two graphs for alleles A1 and A2, and with the two populations plotted on each. This means that the lower graph is the mirror image of the upper one, so you really only need to look at one to see what's going on.

Now, leaving those settings, click in the "Migration?" box and for rate put 0.1 (10%). (The rate values go from 0 to 1.0.)

Run your 10 simulations and note the fixed allele for both populations.

One thing you might notice is that without migration, the allele frequencies of the two populations change independently. They might both up for a while, then one goes down while the other still goes up, and so on. The allele frequencies are changing randomly. What is different when you allow migration?

Play around with the migration rate; try several values, such as 0.01, 0.25, and 0.75. As the rate of migration from one population to the other increases, do the allele frequencies in the populations become more stable or less? Are the populations more similar to each other or less?

Write a little description of your results for these experiments so you'll remember what you found as we move on.

3) Set the population size at 200, A1 allele frequency at 0.5, and now change populations to 3. Run your experiments as before.

Write a few sentences describing how the results with 3 subpopulations are different than 2.

4) In these experiments, we kept the size of each population the same (200). Now let's pretend we have 300 animals and divide them into subpopulations but keep the size of the groups equal. So you will have 1 population of 300, 2 populations of 150, 3 populations of 100, and 6 populations of 50.

We know that diving the population and allowing migration reduces the probability of completely losing an allele and the population is genetically more stable. Now, if you divide the population the simulation is more like the real world because the subpopulations will be smaller.

Run some experiments as before, and fiddle around as you wish with the same settings while thinking about your own breed - its size, subpopulations (perhaps in various countries, different kennels, or different types), and migration rate.

5) Create 3 populations of 100 animals. What is the lowest migration rate that you would need to have to keep from losing one of the A alleles from any of the populations in 200 generations when the initial frequency is 0.5?

6) If you had a population of 300 animals, and you planned to breed in a way that would result in the "migration" of 10% of each subpopulation to another one (doing this via outcrossing), how may subpopulations would you create with your 300 animals to get the best genetic stability in all of them?

7) What did you learn from this? How does the size and number of subpopulations affect genetic stability at a fixed migration rate?

8) Can you think of any "breeder myths" that you just busted with your population simulations?

You made it through this lesson. Did you learn something that changes the way you think about breeding and genetic management (or lack thereof) in your breed?

This was a single unit from a course with 15 modules, each with multiple units (averaging 3-5). Would you benefit from knowing more about how population genetics affects your breeding program? 

You can sign up to take this course here,
​https://www.instituteofcaninebiology.org/openreg-managinggenetics.html

or have a look at the other courses on offer here -
https://www.instituteofcaninebiology.org/openreg-managinggenetics.html

I do recommend that you take Managing Genetics first because it covers foundational topics necessary to understand the material in other courses.

Many breeders stress that the key to improving canine health is DNA testing and selective breeding, and many national kennel clubs and breed clubs endorse this view. However, the evidence does not support the claim that this strategy will prevent genetic disorders or even improve the health of dogs. In fact, it is easy to show why this will not work based on simple principles that any breeder can understand.

Let's say we start with a small population of healthy dogs that are not related to each other. Let's put them on an island and allow them to breed freely for many generations. Over time, individuals in the population will become more and more related to each other because of the limited number of individuals to mate with. Just by chance or through selection,  some alleles in one generation will not get passed on to offspring in the next generation and will be permanently lost from the gene pool. At the same time, unavoidable inbreeding will increase the level of relatedness among the individuals. The net effect is that some individuals will have two copies of a dysfunctional allele at a locus instead of at least one copy of of the normal, "wild-type" allele. 

The replacement of normal alleles with mutations will results in a decline in health and "vigor", including both genetic diseases and measurable traits such as reproductive success, lifespan, and survival. At this point, the population can be considered to be unhealthy, and the problem is the erosion in the quality of the gene pool.

If this was a dog breed displaying some health issues, we would assume the problem is genetics and start looking for a mutation. Sometimes one is identified, but more often the situation appears to be complex and we are unable to identify a clear cause among what we assume is a background of polygenic effects.

We have had only very limited success dealing with problems like this. There is no easy way to select against multiple unknown genes, so usually animals are removed from the gene pool. But this rarely solves the problem. New animals similarly affected continue to appear, sometimes stimulating more research to find the offending mutations. Ultimately, the solution often becomes the removea of animals from the gene pool once again.

After decades of engaging in similar strategies, breeders have not eliminated any of these problems, except where caused by a single recessive or dominant mutation. Why is this? 

We identify the problem as "bad" genes we have to get rid of. Rather, it's not the bad genes that are the problem, it is the LOSS of the normal, functional genes that should be found at that locus. They, of course, are lost through genetic drift and selection with each generation of breeding. If every unique allele in a breed's gene pool was passed on to an animal in the next generation, then the composition of the gene pool would be stable over time, and we should be able to produce healthy, quality animals generation after generation. But we don't breed in a way that protects the contents of the gene pool. Thus alleles are lost every generation, either by design or by chance. We do not protect the genes that are necessary to produce healthy animals, so as they disappear from the gene pool, we have to expect that we will see problems.

Interestingly, most wild animal populations can persist for hundreds of years, maintaining health and function because the complement of genes necessary to produce the animals is constant. How do these unmanaged populations of animals do this? It's simple. The alleles that are lost are replaced when new individuals join from other populations. This is also done deliberately in commercial animal breeding, when a single population is subdivided and managed in a rotational breeding program. This practice restores genes lost from one subpopulation through the introduction of animals from another subpopulation. A carefully run program can continue to breed in a closed gene pool for many generations, simply by preventing the loss of alleles from the gene pool. The valued traits of the animals are preserved and the health and quality of the animals are maintained at a high level. 

Note that the gene pools of ALL animals contain mutations. Why don't these mutations result in the number and magnitude of problems we wrestle with in purebred dogs? The dominant mutations are removed if they have a sufficiently detrimental effect on the animals. Others will only cause issues if paired with another copy of the mutation, because the locus for this gene will have two copies of a non-functional gene. What is absent is a copy of the normal allele necessary for a particular function. 

The problem isn't that the animal has a mutation; it's that it lacks a copy of the normal allele. Simply "getting rid" of the mutation will not solve the problem. The problem is not a malicious genetic bomb in the gene pool. It's the absence of the normal alleles that should be in its place.

What is critical for breeders to understand is that there is only ONE solution to this problem. It is to restore the alleles necessary to perform every function of the body of a dog. DNA testing will spot a few of the "bombs". But throwing those out still leaves a gene pool without the allelic diversity necessary to produce healthy animals. Every breeder knows from experience that trying to remove mutations by removing animals is fruitless; the result will not be a happy, healthy population of dogs.

Can you see that DNA testing to identify the few mutations we know about will not restore health to dog breeds? As long as the gene pool is closed, we cannot restore the gene pool to health and we will not be able to produce healthy dogs sustainably. If we want healthy dogs, we need to produce dogs with the complement of alleles necessary to support all the bits and pieces of function necessary to be a healthy dog. All breeds in a closed gene pool will already be suffering from a loss of alleles, some of which will result in loss of function. So, not only do we need to breed in a way that protects genetic diversity, we must take whatever steps are necessary to restore the diversity of the gene pools.

Our focus on DNA testing and selective breeding is failing to improve the health of dog breeds, and this is why. Advice to breeders that we can produce healthy dogs if we continue to use our current strategies is simply wrong because it suggests that loss of alleles every generation will have no detrimental effects on health and function. This is impossible.

There is lots we don't know about genetics, but there is nothing that would explain how we might continue to breed dogs as we are without the deterioration we are seeing in the health and quality of the dogs. In fact, the health issues that we wrestle with in purebred dogs are caused by the way we breed. To change the trajectory of declining health in dogs, we must change the way we breed.

The solution to our problem is obvious and conceptually simple. It has been practiced by breeders of other animals for many years. There is no magic necessary, and it's not hard. You do need good information and guidance from individuals that understand the details of population genetics in order to accomplish the genetic rehabilitation of a breed effectively and efficiently. The sooner you start, the sooner we can be dog owners and breeders without also being experts in cancer, epilepsy, kidney failure, or sudden cardiac death.

Finally, the average dog lover is growing intolerant of the high incidence of health problems in purebred dogs. They identify the breeders as part of the problem, and they are not wrong. Pressure on breeders to improve health doesn't seem to be resulting in significant improvement, so the impetus now is legislation that will regulate dog breeding and make it illegal to produce animals that can be expected to suffer from health issues. 

Breeders have experience in dog breeding. We must pair this with expertise in the proper genetic management of animal populations. To restore dog breeds to health, we must encourage and support this essential partnership.