Breeding related

It was quite a long and often technical story about the various aspects of genetics. The question comes to mind: what is it good for? For a good practice of breeding, but also for the understanding of breeding issues by puppy buyers, knowledge of genetics is very useful. Inherited diseases are often discussed, but what are the implications for the breeding animals and their offspring? If the half brother of my dog has a disease, what does that mean for my dog? What colours can occur in a litter? These are questions that can be answered when one has knowledge of genetics.

In one of the following articles coat colour genetics of the Border Collie are discussed. The interaction between the various alleles and genes is relatively simple. A number of slightly more complex combinations were already discussed in the articles about interaction between alleles and interaction between genes. At the end of the article there is a small overview of the various genotypes and phenotypes. From this table one can easily determine the genotypes of a certain combination of dogs and from there the possible genotypes of the offspring can be 'calculated'.

There is however a catch: several colours have one or more dashes in the genotypes. This means that the content of that allele is not important for the phenotype; for the colour to become black in a dog (B-allele, dominating the b (brown) allele) it's not important whether the dog is BB or Bb. For the result of a litter, it can be important! In this case we have to study the pedigrees of the ancestors. Especially colours which fully determine the relevant genes are important, but also a study of previous litters can be helpful. You can sometimes determine the complete genotype of the dogs (as far as colours are concerned), but often you have to guess quite a bit if you don't want to end up performing very complicated calculations of probability. From a mating of a blue merle and a black and white BC you will most likely get only blue merle and black and white pups. Recently we learned of litter that had also chocolate and white (=red/white), red merle and tricolour pups. Afterwards you can explain this by the fact that both parents must be 'carriers' of chocolate (red) and tricolour genes. Five colours in one litter is very rare. In this case it gives a good look at the genotypes of the parents.

It's also possible to perform DNA tests for coat colours. Later we will see that this can useful in some cases for health reasons.

In the health articles the inheritance of various diseases is mentioned, if the inheritance is known. If disorders of pups are reported to the breeder, (s)he can assess whether a certain dog is a carrier or not. If it is known that a certain disease is caused by a simple autosomal, recessive gene, we can conclude that both parents of an affected dog must be carriers.

Often little or nothing is known about the inheritance of a disease, so hardly anything can be concluded from affected pups. If the disorder is determined to be inherited, but the mechanism is unknown, a repetition of the combination of dam and sire must be avoided.

A number of diseases is known to have a polygenous inheritance. This means we have to look at the article population genetics (2) about the qualitative characteristics. By collecting information about the offspring we can determine the breeding properties of the parents to some extend. In the case of this kind of disorders we must also pay attention to the heritability index. A low or very low heritability index makes it more likely that big differences from the parents were caused by environmental factors. The example of HD provides a bit of a background:
A pup from parents who both have HD B1 is tested HD C2. With a heritability index of about 0.3 it is unlikely (but not impossible) that the effect can be explained by genetic influences. If we keep in mind that it is a polygenous disorder where many genes have a adding effect (adding positive and negative influences), it is plausible that a certain combination of positive and negative genes shows this enormous deviation. On the other hand there are many environmental factors that can influence the expression: a dog that should genetically be rated HD C1 can become HD C2 because of these factors. Positive environmental influences can make this dog to be rated HD B1.

How can we determine the genetic influences and the environmental factors? This leads us to statistics. The key is to keep one influence constant. We can study a large number of animals in the same environment (from womb till living environment) and thus be looking at the genetic factors. On the other hand we can also raise number of litters from the same parents in very different environments. We can now compare these with the other animals who also live in different environments. If we measure a characteristic we will find a distribution of values with an average. The difference between the average of the entire population and the group we studied must be caused by the genetic factor. This is only true for large numbers of animals. A statistician can calculate for what numbers the measurement is meaningful.

We must make a difference between the type of inheritance.

Simple inheritance

Most of the times this concerns a recessive autosomal gene. In a previous article (Population genetics) we have already studied selection against a recessive allele. A complication is the relatively large number of carriers compared to the small amount of visibly affected animals. A genetic blood test would be useful to detect carriers. We tend to remove all carriers from the breeding stock at that moment. If you take into account the large impact this has on the number of available animals for breeding and the other characteristics of the detected carriers, it could be wise to use certain carriers for breeding. You can now prevent the birth of affected pups by only breeding carriers to dogs that are free of the disease. Slowly the carriers can be removed from the breeding stock and the diseases is removed.

Test breeding

A different, hardly ever used approach to find carriers is the use of test breeding. In reality this is a costly affair with an ethical problem. We will briefly look at the possibilities that can be used in dogs.

1. Mating a suspected dog with homozygous recessive bitches (affected dogs). Provided that the disease is not lethal. Statistical calculations show that you can say with 95% certainty that a dog is free from this disorder if he produces at least five healthy pups. A big disadvantage is that you have to keep bitches especially for this purpose and that you can only test for one disorder at a time.
2. Mating a suspected dog to heterozygous bitches (carriers). This situation requires eleven healthy pups to say with 95% certainty that the dog is free. This usually means two litters. This also has the disadvantage of testing one disease at a time.

A probably bigger problem are the pups that are born from these litters. In situation 1. a full 100% of the pups is carrier and in situation 2. even 50%. The only reasonable solution seems to be to put these pups to sleep, which is not a very humane thing to do.

Pedigree studies

The problem here is that a pedigree does not contain enough data. And even if one could obtain enough data from breed clubs or kennel clubs the fact that a pup is affected by a disorder is often kept a secret by breeders (or at least not widely published). If you happen to find affected dogs (because a mandatory test is performed and the results are registered) you can use the statistical limits stated before to declare other dogs to be free (with 95% certainty). Example:
Let's assume that disease X (recessive autosomal) has the property that by the age of five you can savely declare a dog to be free if he doesn't show symptoms. Dam A is seven years old and was found to be affected. None of her eight pups (who are five years old by now) by sire B are affected. The sire can be assumed to be free (with 95% certainty). This information is very valuable for other litters by the same dog.

DNA tests

In the past decades affordable and relatively quick DNA tests have become available for a lot of inherited diseases and also for coat colours. This is a really powerful weapon in the battle against diseases. The number of available tests has grown to quite an impressive list; there's a separate page with small descriptions.

Simply because we can test an individual and know if they are affected (homozygous recessive in many cases), carrier or free makes it possible to also use carriers in a breeding program. If you can't detect carriers with a DNA test you have to exclude known carriers from breeding because it is not unlikely that a mating would be done with a carrier that isn't known yet. Once you have to possibility to test an individual it is an option to use a carrier and make sure the mating partner is free for that particular disease.

This way you can keep the population as large as possible (fighting inbreeding) and be sure that you won't have pups that are affected by the disease. Breeding healthy pups is huge responsibility for breeders and DNA tests are an important tool for this. If you want to remove a certain disease from your lines you can also test the pups in a litter and select one that is tested as free for your future breeding plans.

Quantitative inheritance

In the previous article we already used CHD as an example to show how hard it can be to select against this type of disease. Because the occurance of the disease cannot be completely explained by genetic influences and because the polygenic inheritance makes a blood test very hard if not impossible to design, selection is all that remains. By only using that part of the population that shows 'good' values for the disorder improvement is certainly possible.

In previous articles we've seen that inbreeding is unavoidable in breeding. With the help of studying pedigrees the amount of inbreeding of a certain combination can be determined. For the rest the use of inbreeding depends on the breeding goal and the comparison of the pro's and the cons. In my oppinion the cons often outweigh the pro's and selection has already enough influence on the genetic variation.

Inbreeding is actually only necessary when one is rebuilding a breed when the breeding base is too smal for 'normal' breeding methods.

A frequently heard argument in favor of using inbreeding or line breeding is that one could fix the properties of a good dog (often male) in his offspring. His daughter gets half of her genes from her father. Crossing father and daughter is supposed to result in a complete copy of the father in half of the pups. A worthy cause that is full of pitfalls:

In the first place a daughter does get half of the genetic material from her father, the question is which half. When cells multiply the genes are distributed in a random way. During the production of reproduction cells in the daughter we're talking about, it's very unlikely that an ovum (egg) gets exactly the genes that were originally from the beloved sire.

Secondly the phenomenon crossing-over is not taken into account. Crossing-over occurs frequently in real life (it's more a rule than an exception), which causes the genetic material of the sire to be spread among different chromosomes (within one pair of chromosomes that is).

In the third place this argument assumes the sire to be (almost) completely homozygous for all properties. This is not a realistic situation. About the characteristics for which this dog is heterozygous the question can be asked: which half is passed on?

In the fourth place many desired properties such as health and working abilities are genetically very complicated. Heterozygous animals are probably in favour as far as these properties are concerned. Actually, with working ability the extremes are often not wanted at all.

The disadvantages of inbreeding / line breeding such as inbreeding depression (reduced health and fertility) are often bigger than the advantages. Especially when selection methods such as mentioned in the previous article can also result in good progression.

Except for the properties of the animal itself, we will need to look at the close relatives for a number of characteristics. We are actually trying to determine the breeding value of the dog. In other words: we try to estimate what the various characteristics of the puppies will be. A good estimation are previous litters. Fertility can be judged by the size of the litter and possible birth problems. If this is not possible we can look at litters from their parents and (half) brothers and sisters.

Concerning the properties that are visible in a dog or bitch, we often look at the characteristics of the animals itself. If possible one would better take a look at previous litters. For the future litter it's obviously more valuable that a dog has already produced a few champion sheepdogs than the fact that the dog itself is a champion. The same applies to disorders such as CHD.

Since it is often impossible to collect data from family members, the choice is limited to the properties of the breeding animals themselves. A breed club could be of help here by collecting as much data from as much animals as possible. It would come handy to take a look at the data from future breeding animals this way. Privacy laws may prevent this kind of data to be available to the public.

A population that is too small means there is a lot of inbreeding. It is clear that quite a bit of the genetic variation is lost by this inbreeding. We need to keep the population large enough to let it survive.

Another threat is using one particular dog too often. This is often a dog that has popular characteristics or a dog that is unrelated to the rest of the population ("fresh blood"). Many a breeder uses this one dog and the next generation contains many related animals. So we also need to prevent breeders overusing one particular dog.

The minimum number of breeding animals one needs to maintain a healthy population is fairly simple to calculate:

First of all we have to see how much inbreeding is caused by which numbers of breeding animals. Wright found an approximation for this.
$$\Delta F = {1 \over ( 8 \times N_m )} + {1 \over ( 8 \times N_f \,)}$$
\(\Delta F\) : increase of inbreeding coefficient
\(N_m\) : number of stud dogs
\(N_f\) : number of brood bitches

In practice an increase of the average inbreeding coefficient of 1% per generation is just acceptible. The counter forces such as mutation and migration are in balance with the influence of the inbreeding.

In dogs the generation interval is about 2.5 years. If we adjust the number of litters within a 2.5 year period for the bitches that had more than one litter during that period and for the bitches that are more related than average, we can fill in the numbers.

\(\Delta F = 0.01\); \(N_f\) has been estimated and with a bit of calculation we can easily find the number of stud dogs that we need per generation. Often this number has to be increased a bit because not all stud dogs are unrelated.

By using the number of litters per year we can also calculate the number of bitches that one dog may serve per year.
Example

There are 100 litters in a year. That's 250 litters per generation. Let's adjust this number to 200 for the bitches that have more than one litter per generation.
Fill in the numbers:
\(0.01 = {1 \over ( 8 \times 200 )} + {1 \over ( 8 \times N_m )}\)
\(0.01 = {1 \over ( 1600 )} + {1 \over ( 8 \times N_m )}\)
\(0.01 = 0.000625 + {1 \over ( 8 \times N_m )}\)
\(0.01 - 0.000625 = {1 \over ( 8 \times N_m )}\)
\(0.009375 = {1 \over ( 8 \times Nm )}\)
\({1 \over 0.009375} = 8 \times N_m\)
\(106.6666 = 8 \times N_m\)
\(13.3333 = N_m\)

This means we need at least 14 unrelated stud dogs to keep the population healthy. For 100 litters per year, this comes to 100 / 14 = 7.14 matings per dog per year