Breed to Avoid Genetic Diseases Picture

English Czech German Danish Spanish Finnish French Italian Dutch Portuguese Swedish Taiwanese

By Ulrika Olsson

The risk of genetic diseases is something that is discussed more and more often among cat breeders and in the different cat clubs and associations. Of course we all want our cats to be healthy! Still bad genetic diseases show up every now and then in different breeds. Why does this happen? What can we do about it? In order to understand what is happening in our breeds when the genetic "epidemics" show up, and in trying to control them, one needs to have a little knowledge about the basics of population genetics.

An important word: GENE FREQUENCY

Using the gene for dilute color as an example, let us assume that we have a breed population of 100 cats. Since every cat has a double set of chromosomes, this population will have 200 loci for the dilution gene - meaning 200 places where the D or d gene can be situated. Now assume that 40 of these loci are filled with a d-gene, while the remaining 160 are filled with D-genes. Then the gene frequency for d in this population is 40/200 = 0,20 = 20%. In the same way we find that the gene frequency for D is 160/200 = 0,80 = 80%.

Now if certain equilibrium criteria (Hardy-Weinberg's equilibrium) are fullfilled - which they never are completely in reality, but let's not bother about that for the moment - then the probability of two d genes to meet in an individual, and result in a blue kitten, will be 20% x 20% = 0,20 x 0,20 = 0,04 = 4%.

The probability for a kitten in this population to be a carrier (heterozygous) of dilute is 20% x 80% + 80% x 20% (the probability of a d gene in one particular of the two chromosomes x the probability of a D gene in the other chromosome + the probability of a d gene in "the other" chromosome x the probability of a D gene in "the first" chromosome) = 2 x (0,20 x 0,80) = 0,32 = 32%.

The probability for a kitten in this population to be homozygous black (non-diluted), DD, is 80% x 80% = 0,80 x 0,80 = 0,64 = 64%.

Now we look at it from the opposite direction. Assume that we have discovered that 16% of the population has PRA. If we could now assume that the above mentioned equilibrium criteria are fullfilled (and assuming among other things that the population is not divided into different lines), then we could estimate the gene frequency of the recessive PRA gene. Let us call this gene frequency f. Then we know that f x f = 0,16. Then we simply solve the equation. f = the square root of 0,16. f = 0,40 = 40%.

We can now also estimate the proportion of heterozygous carriers of the PRA gene: (0,40 x 0,60) + (0,60 x 0,40) = 2 x (0,40 x 0,60) = 48%. Very handy, don't you agree? Now over to the subject of gene pool and inbreeding, a very important subject!

Effective Population

The fact that some breeds are larger in population than others and also the fact that the gene pool can be too limited in a small breed is probably obvious to most breeders. But far too many breeders feel secure in that the larger breeds - for instance Persians and Birmans - have a large enough gene pool. No risk of inbreeding problems there, unless you choose to deliberately inbreed a line! But that is not always true.

We sometimes discuss the fact that overbreeding from a few outstanding individuals is harmful for breeds, even the larger breeds. The most extreme overbreeding from one individual that we could imagine is if one single male is mating every female in a population in a generation. With that kind of breeding, of course the gene pool will not be large, even if the population contained a thousand unrelated females. In order to get a better idea about just HOW large the gene pool really is in such a case, we could calculate the effective population. If the population contains 100 individuals, with an equal number of males and females, all of which are mated randomly to each other with the same amount of offsprings resulting from every parent, then the EFFECTIVE population is also 100. On the other hand, with the extreme overbreeding of one male described above, we can calculate the effective population with the following formula:

       1        1        1

      ---- = ------ + -------

       Ne    4 x Nm    4 x Nf


Where Ne = the effective population, Nm = number of males, Nf = number of females.

In our case:

       1       1         1

      ---- = ----- + --------

       Ne    4 x 1   4 x 1000

       1      1     1

      ---- = --- + ----

       Ne     4    4000

       1     1001

      ---- = ----

       Ne    4000

       Ne    4000

      ---- = ----

       1     1001

      Ne = 3,996

So this large population - 1001 animals used for breeding - equals a population of only 2 males and 2 females!

Generally speaking, the effective population will not be larger than 4 times the number of individuals of the least represented sex. (Unless you have set up a breeding program specificly designed to avoid losing genetic variation, but that is very seldom the case in cat breeding.) This means that if 5 males are used, the effective population will not be more than 4 x 5 = 20, even if we could use a million different females in that breeding program.

In reality of course it is seldom the case that five males are used for equally many litters and the remaining males in the population are not used for breeding AT ALL. Then it gets a bit difficult to use this formula. But don't worry! There are other methods!

A too small effective population will cause the degree of inbreeding to increase for every generation. There is a connection between inbreeding and effective population. Using this connection we will be able to calculate the effective population of our cat breeds. We can use our pedigrees to calculate the coefficient of inbreeding, COI. The easiest way of doing this is to enter the pedigree in a good pedigree program that is capable of calculating the COI. It is also possible to calculate it by hand. It is not difficult at all, but if the relationships are complicated and you want to calculate the COI from many generations, it is very time consuming, and there is a considerable risk to make a small error somewhere along the way. However, if the relationships are simple and the number of generations to calculate from is reasonable, you could quickly calculate the COI directly from the pedigree. It really is a good thing to know how to do! There have been many articles published devoted to explaining how.

If we would like to calculate the effective population of a particular cat breed in a specific country, for instance Cornish Rex, we would first need a file or a database with the pedigrees of the Cornish Rexes in the chosen country. From this database we pick out the litters that were born during the latest year. If there are very many litters, a sample with the desired number of litters could be selected randomly from those litters. Then we calculate the COI for these litters, for instance calculated from 5 generations. Take mean value of these COIs, and keep it for later use. Then we take all the parents of our selected litters and calculate the COIs for those cats, only now we calculate from one generation less - in our case 4 generations. And then we take the mean value of these COIs. Assume that we got the mean value 5,5% for the litters and the mean value 4,5% for the parents of the litters. We then take the difference between those two numbers, 5,5% - 4,5% = 1,0 percent unit. We are interested in the percent with which the degree of heterozygosity has been decreased in this generation, and that will then be 0,01/(1-0,045) = 0,01047, or 1, 047%. And then we have another formula:


      Ne = ------

           2 x dF

... where Ne is the effective population, and dF is the decrease of heterozygosity in the latest generation, as calculated above.

In our example, with dF = 1,047% = 0,01047, we get:


      Ne = ----------- = 47,76

           2 x 0,01047

The effective population of Cornish Rex would then in our example be 48. An effective population of 50 is generally considered to be the limit for when a species is considered to be directly threatened by extinction. Now of course our cat breeds are not separate species, but with closed studbooks the effect will be the same.

Now it could be the case that the Cornish Rexes in our population are inbred to a greater extent than what is really necessary, only because the breeders are separated in smaller groups, breeding the cats in different lines. Perhaps our effective population could be larger if this segregation into lines didn't exist? How can we check this?

Then we go back tho the group of parents of the litters in our sample, divide them into males and females, and then we test mate them randomly. The COIs are calculated for those random matings, we calculate the mean value of them, and then compare this value with the mean value of the COIs of the parents again. Assume that the random matings had the mean value of 5,0%. Then dF = (0,05-0,045)/(1-0,045) = 0,005236 = 0,5236%. And...


      Ne = ------------ = 95,49

           2 x 0,005236

So, in this case the effective population will be at a much more reasonable level if we just can achieve a better cooperation between the breeders in the country.

What happens if the effective population is too small?

One thing that happens is that the inbreeding coefficient increases for every generation. Actually, this happens for ALL populations that are not of unlimited size, but then the natural selection is probably working against more inbred individuals, so that reasonably small increases of the inbreeding are adjusted and status quo is kept. It is also known that a larger number of eggcells get fertilized than the number of offsprings that finally are born in the litters, and one theory is that these early featuses have to "fight" for their place in the uterus, and that the more homozygous featuses are less likely to survive. This theory however not been proven to be true.

What will happen, when the inbreeding coefficient is increased generation by generation? In the beginning, not much at all. It isn't until the degree of homozygosity reaches a certain critical level that the real problems show up, and then it is usually much more difficult to take corrective action. It is much better to start working against these problems before the symptoms show up. The pedagogic problem then of course is that those who begin to breed with too few individuals won't immediately see the problems that it will cause. They think "I have bred in this way for many years, and I have not had any problems". But as we can see, "trial and error" is not a very good approach here! When "error" appears it is a bit late to adjust in an easy way.

Why then is inbreeding so dangerous? One thing - that every educated breeder knows about - is that it means an increased risk of doubling up on harmful or lethal recessive genes. The double set of chromosomes otherwise protects us from this risk, to a great extent in a population that is not too inbred. All individuals carry a few harmful recessive genes. Some people think that the inbreeding clears out the harmful recessives and leads to a healthier breed for the future. But first, the inbreeding doesn't clear out anything in itself, it has to be combined with a strong selection in order to clear out any undesirable genes. Second, you have to inbreed incredibly strongly in order to get all or almost all loci homozygous, so that you can see what the cats are carrying and weed out all the undesirable genes. Mate a female with her full brother, and 25% of all loci are homozygous. Then mate two of these offsprings with eachother, and 37,5% of their loci are homozygous. And then we take two of THOSE offsprings and mate them with each other! Now the inbreeding is so strong that most breeders would back away. But still "only" 50% of the loci are homozygous. So in spite of this drastic inbreeding, we will miss out on exposing other recessive, potentially harmful genes.

But let us assume that we go all the way with this! We breed a line towards 100% homozygosity, selecting strongly against harmful genes all the way. All individuals will then have exactly the same genotype, except for the fact that the males must have a Y chromosome where the females have a second X chromosome.

OK, it took a lot of effort and money to make this so called isogene line, and many cats died on the way. But if we now finally have reached to this point, then we have a line that is 100% healthy from a genetic point of view! Yippeeee!!!

It is possible, this can be done, if you are careful not to let the level of homozygosity increase more quickly than you manage to weed out the bad genes. This has been done with mice that are to be used for scientific tests. It works very well! But... they only manage to get about one line out of twenty to survive. The other 19 lines are dying in the process. Maybe better not to take that chance?

Also, the immune system is not very good in homozygous individuals. The immune system works much better if the loci involved are heterozygous, since this gives the individual the possibility to develop more DIFFERENT kinds of antibodies, not just lots of antibodies of the SAME kind. This is not a major problem in laboratory mice, since their environment is quite protected from (undesired) contagious diseases, and since it isn't exactly considered to be a tragedy, unfortunately, if a laboratory mouse dies. If on the other hand a much loved pet cat and family member dies it is indeed very sad. Hmmm... Perhaps not such a good idea after all?!

On top of that mutations happen spontaniously and would with time destroy our fine genotype. You have to count on one or two new mutations in every individual.

I think we had better change our strategy!

But what if the breed is already inbred?

If a breed or a population is already so inbred that clear signs of inbreeding depression have shown up, for instance a high rate of early cancer or infections, what to do then?

If there are unrelated lines in other countries, of course the best solution would be to increase the exchange of cats between those countries. If such unrelated lines are not available, we will have to outcross to another breed or unregistered cats that fit the standard reasonably well. If enough new genes are mixed into the population, the inbreeding problem will be solved.

A not uncommon objection against these kinds of solutions is that we don't know which new harmful recessives might be introduced into our breed through these outcrosses. That is true, we don't know. What we do know though is that most individuals carry some harmful recessive genes. Many breeders also think that it is better to have a more inbred population less different kinds of genetic diseases, in order to more easily keep them under control. Maybe there are even tests available for those deseases. But, as we will see below, it is better to have lower frequencies of several different harmful recessives than to have a higher frequency of one single recessive.

Assume that we have a population A with a gene frequency of 50% for some kind of recessive defect. We will compare it to a population B with gene frequencies of 10% for five different recessive defects. Both populations will then have the same frequency of harmful genes, but population A has defect genes of only one kind (easy to keep under control) while population B has its defect genes divided in five different kinds.

The risk for a kitten in population A to show the genetic disease is then 0,50 x 0,50 = 0,25 = 25%.

The risk for a kitten to show a genetic disease in population B is 5 x (0,10 x 0,10) = 0,05 = 5%. This shows that we get considerably less defective kittens in a population which has lower frequencies for several different kinds of diseases. The most effective way to keep a breed healthy is not to try to eliminate the harmful recessives, but to get the frequency down to a such a low level that two harmful recessives of the same kind almost never meet.

Some breeders will hesitate to outcross because they are afraid that the type will be gone for ever. Some breeders are of the opinion that inbreeding (linebreeding) is the only way to get excellent and uniform type. It is true that by using inbreeding you can achieve quicker results in this area. The problem is that you risk the long term health of the cats. It is possible to achieve the same result without inbreeding, although it takes more time. Unfortunately inbreeding is a very tempting shortcut for breeders that are interested in showing their cats. But one should keep in mind that most of the genes that are doubled up by inbreeding have absolutely nothing to do with type. For instance, a human has approximately 100 000 genes, and 98,5% of those are identical with those of a chimpanzee! And still, aren't we rather different from a chimpanzee? How large a part of the genes could different between a Siamese and a Persian? Or a Norwegian Forestcat and a Maine Coon? Or between a Burmese of good type and a Burmese look alike moggie with reasonably good type? Not more than what we can fix with some generations of selective breeding, I'm quite sure about that!

Genetic problems are showing up in the breeds!

Yes, that is also due to a too small effective population! Unless it is due to a breeding without regard to the anatomic functions of the animal. Breeding for extremely long bodies might cause problems with the back, and breeding for very short faces might give problems with the teeth, breeding for extremely triangular, square, round, etc. heads might cause problems with jaws, eyes, brain, or whatever. A cat must be allowed to first and foremost be a CAT. It is not a piece of clay that we can shape after our own estetic ideals. A cat doesn't consist of circles, triangles, squares, or other geometric figures, we have to remember that. Sometimes I tend to agree with a genetiscist, mostly working with dogs, who suggested that perhaps we should breed all dogs with the coats of a poodle, so that we could CUT out the geometric shapes and odd angles that we find attractive. Then the animals could have their anatomy in peace.

Except for this sad phenomenon, it is the too small effective populations that cause the high frequencies of many genetic diseases to show up in breeds. Many breeders seem to be a bit confused about this. They might think that if we have for instance 10% of the cats in a breed affected by PRA, meaning a gene frequency of approximately 32% for the recessive PRA gene, and if we don't test and work to reduce this frequency, then the frequency will automaticly increase with time. This is of course not correct. If it were, then also the frequency of diluted (blue, cream, etc.) cats would increase all the time unless we selected against the dilution gene. If the effective population is big enough, and no selection for or against PRA is made, then the gene frequency will stay at 32%.

On the other hand if we select ever so weakly against PRA, for instance let the cats who themselves have PRA (homozygots) have no more than one litter, then the gene frequency will decrease. Slowly with a weak selection, more quickly with a strong selection. If we could now remove all cats showing PRA from the breeding program before they have any kittens at all, the result would be this:

      generation      Frequency of genotypes (%)

      number          PP         Pp        pp

      1               46,8       43,2      10,0

      2               57,8       36,5      5,76

      3               65,0       31,2      3,74

      4               70,2       27,2      2,62

      5               74,1       24,0      1,94

      And so on.

But then what happens if the effective population is too small? What will then happen with the gene frequency? It will be the same effect as if you toss up a coin 10 times. Your chance for heads is 50% every time. And if you had tossed up this coin 1000 times, you would have got quite close to 50% heads and 50% tails. But now you only toss it 10 times. Then it is not very surprising if you by chance got 70% heads and 30% tails, or 30% heads and 70% tails, or something like that.

In the analogous scenario in a small cat population, it means that the gene frequency of around 30% in the next generation might have increased to 35%, because of this random effect. Or else it might have decreased to 25%, because of the same random effect, which in the PRA case would of course be a lot nicer. But let's be pessimistic and assume that the frequency increased to 35%. Then the EXPECTED value of the gene frequency for the next generation is also 35%. But by chance it might end up at 29%, 34%, 38%, 42%, or whatever. The smaller the effective population is the larger the risk is of getting a large deviation from the expected value of the gene frequency. Then this frequency, that we got randomly, will be the expected value for the next generation. This phenomenon is called random drift. If the impact of this random drift gets stronger than the impact of the selection - natural or artificial - then the changes of the gene frequency could very well be the opposite to what we wanted. DESPITE the selection. Then the eyes might get paler in our Siameses, or the lynx tufts might get smaller in our Norwegian Forestcats, or PKD might get more common in our Persians. That would of course be anything but funny!

If we now look into why PKD became a far too common problem in the Persians, it could hardly be caused by some mysterious selection in favour of lumpy kidneys. It must have another cause.

It must of course have started with a mutation in a cat long ago. It was a dominant gene, so the cat developed cycts on its kidneys. Let us assume that it was a male that died from the PKD at 5 years of age. Or perhaps at 7-8 years of age. Either way, we have a certain selection against the gene. If the population is then large enough, the frequency will then decrease and eventually go down to 0%. And even if there were no selection against the gene what so ever, there would be a good chance for the gene to disappear within a few generations, since the frequency could by chance be slightly larger or slightly smaller. And since the frequency was initially very small (one mutated gene in a large population), it is rather likely that the frequency by chance happened to decrease to 0%, and then the gene is gone.

So, the effective population for the Persians apparently was not large enough. Random drift resulted, and by chance this unfortunately caused an increase of the frequency of the PKD gene. In spite of a certain amount of selection against the gene, the result was that the frequency ended up at approximately 25-30% before more breeders became aware of the problem and a stronger selection was introduced.

What does all this tell us? That if we don't have large enough effective populations, then high frequencies of unpleasant genetic problems will continue to pop up. If we are unlucky we might also have difficulties attempting to reduce these problems with selection.

If we instead make sure that we have large enough effective populations in our breeds, genetic diseases will not pop up as a common problem in the entire population. And as a bonus we avoid inbreeding depressions and bad immune systems.

To breed with too small effective populations, and at the same time start projects to fight against genetic diseases within a breed, is just like being treated for lung cancer and continue smoking. Or to scoop out and wipe away the water that have poured over the rim of the bath-tub, while we still haven't remembered to close the tap, so the water is still pouring INTO the tub.

To work for large enough genepools is a kind of preventive medicine measures for the breeds. It doesn't seem clever just to treat the ailments, without bothering about the preventive measures that could prevent the ailments to arise in the first place.

We should also keep this need for reasonably large effective populations in mind when we create and accept one spotted shorthaired cat breed after the other, and one large furry semilonghair breed of medium head-type after the other, etc. Unless the number of breeders willing to work with spotted shorthaired breeds is increased at the same rate as the increase of the number of breeds, the reqruitment of for instance Bengal breeders - just an example - will be at the expense of the Ocicats', Spotted Oriental Shorthairs', Egyptian Maus', etc. possibilities to keep a constant and large enough number of animals for breeding in their breeding programs. Can these breeds afford this? Will the new breed be able to create a place for itself among all these other breeds? Or perhaps they will all get too small populations, so that we in the end have destroyed all the spotted shorthair breeds? These are important things for us in the cat organisations to think about. These threats are real, not just "in theory", and we have already begun to see the first effects, although not yet as badly as in the dog breeds. But I'm afraid it will get worse, unless we do something about it - the sooner the better.

So this is something that we in the cat organisations have to start working with. It must not be forgotten among all the specific projects against specific diseases. It is about the actual basis for the health of the cats.

©Ulrika Olsson, 2002

Many thanks to Cynthia Bowen, Coontopia cattery, for help with the complicated English language!

Databases on the Internet

In order to be sure that the effective populations and genepools are large enough for the breeds that we are working with, it is necessary to create databases with all cats gathered, regardless of what associations they are registered with. The population genetiscists will then have an opportunity to study the situation for every breed, and we could get breeding recommendations based on facts instead of guesses about what we THINK is the situation. We could also get to know in time, before the SYMPTOMS of inbreeding show up, if the breed is on its way to lose genetic variation in its genepool.

In order to inform and support the breeders worldwide, it is also important that these databases are available on the Internet, so that everyone can share the information that has been gathered. As I see it, the registrations are made in order to benefit the breeding, and then the information also must be available to the breeders. After all, it is the breeders that are making the breeding decisions. So the registrated information must be released for breeders all over the world to study.

Initially, when a more open policy is being discussed, it is common that some people are worried that the release of information will make it easier to falsify pedigrees. But the question is if it doesn't rather make it more difficult to falsify pedigrees, since also those pedigrees soon would show up online. And then the owners of the cats involved would of course react immediately and actions could be taken to correct what is wrong.

All this could possibly sound a bit scary to someone who hasn't been thinking about it before, but I believe that the attitude that the pedigree information should be kept secret is only a question of tradition. As I see it there would be many benefits if the registrated information was released - one just has to get used to the thought of it and convince oneself that nothing dangerous is going to happen.

I am firmly convinced that "bad" breeding choices, from a health point of view, to the greatest part is caused by lack of information and knowledge about the mechanisms. These databases could, combined with increased educations organised by the cat clubs, very much contribute to increased knowledge about the breeds and breeding among the breeders. That would then lead to improved genetic health in our cats.