Reducing genetic risk (does not include edits)

Many canine genetic disorders are caused by autosomal recessive mutations. A dog with only one copy is usually unaffected if the other copy of the gene is the normal version.  But if two animals that both have one copy of the mutation are bred, there is one chance in four that an offspring will be affected because it receives two copies of the mutation. This is from the simple Punnett square - 25% of the offspring will get two copies of the normal gene, 50% will be carriers, and 25% will be homozygous for the mutation.

With the availability of DNA tests for many genetic disorders caused by autosomal recessive mutations, breeders can determine the carrier status of a dog and completely eliminate the risk of producing affected animals. Because these DNA tests reduce the potential risk of producing an affected dog from 25% to 0%, they are a very powerful way to manage specific genetic disorders in purebred dogs.

Unfortunately we have DNA tests for only a fraction of the known mutations in dogs, and there are probably hundreds more recessive mutations lurking in the gene pools of our dogs that we know nothing about. It's easy to see how we can manage the risk of problems from recessive mutations that we can test for, but what can we do about all the ones we don't?

This is a really important question, and one we don't often consider. The question we should ask first is whether we really need to worry about those unknowns. We don't know what they are or what problems they might cause (and of course, a mutation might be "neutral" and not cause a problem at all), so how big of a problem could they be? Consider that the mutations that we now have tests for only showed up on the radar because people starting seeing dogs affected by a genetic disorder. Before this, the mutation could have been passed from parent to offspring without harm for dozens or even hundreds of generations. As a general rule of thumb, we can expect every animal to carry a few mutations that would be lethal in a homozygous offspring, and many more will have some significant impact on health. So there are no doubt many more recessive mutations in the genomes of our dogs that will surface and cause problems if puppies get two copies.

If we can agree that we should assume that the genome of every dog will carry some mutations we don't know about that have the potential of causing trouble, then it is important for us to consider the risk not just from known mutations, but from those unknown mutations as well. We know that we can reduce the risk of affected offspring to zero with a simple DNA test, and that is certainly worth doing. But what can we do about the mutations we can't test for? Can we do better than just crossing our fingers and hoping?

Fortunately, we can. We understand how recessive mutations work. Puppies that get two copies will be affected, so our goal should be to find a way to reduce the risk of breeding two carriers. (There is no risk from breeding a carrier with a dog that is clear of the mutation.) But since we can't know who carriers are, or how many there might be in the population, how can we approach this problem?

Let's start with an easy case, the problem of the dominant mutation. We're going to walk through a series of simple graphs, so just follow along. (They're easy, I promise!) We know that a dog only needs to inherit one copy of a dominant mutation to be affected, so a carrier bred to another carrier or a clear dog will both produce 100% affected offspring. I show this in this graph, where coefficient of inbreeding (COI) is on the x axis and the risk of producing affected offspring is on the y axis. Remember that COI is the probability of a dog inheriting two copies of an allele from an ancestor on both sides of the pedigree. Because you only need to inherit one copy of a dominant mutation to be affected, you are just as affected when COI is zero as when it is 100%. If one parent has the mutation, risk is always 100%.

 

What if the mutation is recessive? In this case, a puppy that gets two copies of the mutation will be affected, so the risk of producing affected puppies is equal to the probability of inheriting two copies of the allele, which is the same as the coefficient of inbreeding. As in this graph, the risk is therefore directly proportional to COI.

We know that if we breed two carriers, the risk of affected offspring is 25%. If we can do the DNA test for that mutation, we don't need to use COI to estimate risk; whatever the COI, we can know the specific risk from that mutation from our test. We can breed a litter with a COI of 10%, or 25%, or 50%, or whatever - and know absolutely that the risk of producing affected puppies is zero as long as only one parent carries the mutation.  So in this case, our graph looks like this:

So this is great. If we have a DNA test, we can reduce the risk of affected puppies from 25% to zero.

Now, how do we deal with all of the mutations we don't have tests for? These will behave as we described above - two carriers will produce puppies with a COI of 25%, so their risk of producing affected puppies is also 25%. In fact, we know that the COI of the puppy (the probability of homozygosity) will be the same as the risk of being affected. The higher the COI, the greater the risk of producing puppies affected by some genetic disorder caused by one of the recessive mutations that we know nothing about.

If you are DNA testing your dogs (and you should be!), you are reducing the risk of a known genetic disorder from 25% to zero. But that doesn't affect the risk from unknown mutations. Presumably, if a new mutation is discovered and a DNA test is available, you would see the value in testing for that as well if there is any question that it might be in your lines.

A responsible breeder can't know the risk of producing affected puppies from an unknown mutation, but they CAN manage that risk, and the rule is simple:

Risk goes down when COI goes down, for both the known and the unknown mutations.

This means that you can reduce the risk of problems from any and ALL recessive mutations by reducing the COI of the litters you produce. Instead of adding each new DNA test to the expense of health testing your dogs, you can effectively manage the risk of genetic disease by managing the risk of homozygosity, using COI as the yardstick.

You will argue that COI is a probability, not a known value. You're right; we can estimate COI and be confident that with a large sample size, the average COI will be about what we predict, but probably not exactly except when sample size is very large. But apart from doing DNA analysis, this statistical estimate of inbreeding is the very best way to predict the outcome of a breeding. Note that your estimate of COI will only be as good as your pedigree database. It needs to go back as far as possible, preferably to founders (and this is why), and missing data will result in an underestimate of the actual COI value, to making the data as complete as possible is important. But with all the caveats, COI is still an extremely powerful tool in the hands of a breeder that appreciates the benefit of reducing the risk of genetic disease caused by any mutation, not just the ones we can test for. And even though the measure of benefit might just be in the reduction of genetic diseases, there are also very real benefits in terms of fertility and litter size, puppy survival and growth, lifespan, and general good health (see some data about benefits of reducing inbreeding here).

If you're paying to do DNA testing, don't turn around a breed a litter with a COI of 25% or more. You will have paid to swap a known risk with an unknown one. If the known risk was worth the cost of a test, consider also the value of reducing the unknown risks. Larger litters of healthier, more fertile, and longer-lived puppies can be had simply by reducing COI. That's a great payoff, and it won't cost you a thing.