By Carol Beuchat PhD

I have discussed here the UK Kennel Club's list of vulnerable native breeds, which is based on registration records. You should read that before continuing here, so you understand what I'm doing.

One problem with their list is that it considers only the populations of dogs in the English and Irish KC's records. It might be the case that the global population of a breed is sufficient to breed sustainably for the forseeable future. To explore this, I have have combined the most recent registration statistics of the UK Kennel Club (from 2014), and the last records made available by the AKC, which are from 2005. (Shame on you, AKC. We need those data!) This still leaves out the registrations in all the other countries, but for now this is the best we can do. (I don't have the Irish KC records that were included in determining vulnerable status by the UK KC.)

I've worked up the data for the Terriers so far, which I present below. In the first graph, I have excluded the data for the Miniature Schnauzer, because the number of registrations is so much higher than the other breeds that many don't even show up on the y-axis scale (24,000 in the US alone). I have also graphed these data with a maximum y value of 1000 in the second graph so the quantities for the breeds with lower registration numbers can be read.

In my earlier post referenced above, I converted registration numbers such as these to effective population size (Ne), which is a standard statistic used to quantify genetic diversity, sort of the "genetic" size of a population. (I explain this in that previous post, so if you skipped reading it and have no idea what I'm talking about, you should read it now.)

Below are two graphs. The top one is for all terrier breeds in the database. A red horizontal line identifies Ne = 1,000, the recommended minimum population size for long-tern sustainability from the 100/1,000 rule (formerly 50/500, as discussed in that previous article). There are two values of Ne for each breed. One assumes that the ratio of males to females in the population is 1:1 (red bars) and the other assumes that only half as many males as females are bred (blue bars). In most breeds, fewer males are used than females, but the ratio can vary widely from breed to breed, and many won't be this extreme.

It is important to note that these data for Ne are almost surely overestimates of the true value because they are based only on the number of males and females that are bred. Other things also affect Ne, such as the degree of inbreeding, which is assumed here to be 0% and is undoubtably 15% or more for most of these breeds (see below). It wouldn't be unreasonable to consider these estimated effective population sizes to be double or more than the actual values.

So, if we take Ne = 1,000 as the minimum sustainable population size for these breeds, then there are clearly many breeds in trouble. Although the Ne for the largest breeds (e.g., Mini Schnauzers, Westies) would seem more than sufficient, genetic disorders are already common in both so there's not much room for optimism that they are safely out of danger because of large population size.

In the lower graph, I have replotted the same data with a maximum Ne of 1,500 to display the smallest breeds more appropriately, and I have added a horizontal line (blue) marking Ne = 100, considered the minimum number of animals needed to found a breed without incurring immediate inbreeding. There are a fair number of breeds that reach the Ne = 100 mark, but remember again that these are surely very optimistic estimates of Ne. If the ratio of breeding males to females is one to one, only two breeds (Bull Terriers and Airedales) exceed the Ne = 1,000 mark, and none make the mark for the lower ratio of males to females.

In the graph above, the effective population sizes of Dandie Dinmonts, Sealyhams, Glen of Imaal Terriers, and Skyes are far below the minimum population size. These breeds have been down there for many generations already and they might hang in there for many more, but if we're going to manage genetic disorders and keep them around for a while, we should be working on a plan now.

For most terriers, the numbers of dogs being bred are simply to low to avoid a steady and inexorable rise in the level of inbreeding, and of course with that necessarily comes inbreeding depression and increased expression of deleterious recessive mutations.

So the prudent question to ask now is about the level of inbreeding in these breeds. Of course, AKC does not provide the data necessary to do these calculations, but the UK KC does offer Mate Select. Their statistics are based on the dogs registered in the previous year (not an average of the entire breed), and their database only goes as far back as their digital records (sometime in the 1960s; why this is bad) and would also not include ancestor information for imported dogs past the previous 3 generations. So in fact, these data really aren't all that useful except as an indication of the rosiest possible picture, because all of these issues will result in an underestimation of the degree of inbreeding. That said, here is a graph of those data.

Let me identify a few critical landmarks on this graph for you. A COI of 6.25% is what you get from the mating of first cousins, 12.5% is the result of mating half-sibs, and mating full sibs or a parent with offspring produces a litter with a COI of 25%.

So what we have are a bunch of breeds with a level of inbreeding AT LEAST as high as what you would get from mating first cousins. In fact, most of these breeds are (optimistically) as inbred as the offspring of a half-sib cross. I noted that these figures would be underestimates; do we have any better estimates to compare them with?

I just analyzed a very incomplete pedigree database for Westies - missing data, misspelled names, the works, so again the COI would be an underestimate, but the data did go back to the turn of the century. From that, the average COI of the current population of dogs (in the world) is about 15%, and I am certain that is a significant underestimate. By comparison, Westies in Mate Select smell like roses with a COI of only 5.6%, so adjust your perception of the data for the other breeds accordingly. I would not be at all surprised to find the true COI for many terrier breeds to be in excess of 25%, but right now we don't have the data to know. Note that a COI of 25% means that 25% of the genes in the dog are homozygous. There aren't enough DNA tests to even come close to identifying all potential deleterious alleles in a dog, and in any case if the consequence of testing is to eliminate dogs from the breeding pool, the inbreeding situation will only get worse.

I have already looked at the historical registration figures for the UK Kennel club, and quite apart from the single-year figures examined here, there are some signs of even more potential trouble. Large fluctuations in population size are very damaging to the gene pool of a breed, as I explained here, and a few of the UK breed populations have just passed or are right in the middle of some major cycles. Westies, Staffordshire Bull Terriers, and Border Terriers have had spectacular increases in population size followed by crashes in the last decade (those graphs are here), so the registration numbers we used for these breeds in our calculations of Ne above were far higher than the more stable population sizes before the boom. The crash of the Westies population is pretty much complete, but the Border Terrier is still on the way down, so there might be some opportunity there to prevent less popular lines or founder genes from being lost if only the data were available to do the appropriate analyses. If there is anyone out there in Border Terriers, it would be worth looking into the situation and mitigate now before the damage is completely done. The Staffie is a little further along, but intervention is possible and better late than never.

Sadly, there is much more to the terrier situation than was revealed in a simple list of "vulnerable" breeds. There are many breeds for which I suspect the breeders are quite unaware of how dire the situation is. New genetic disorders pop up regularly, litter sizes get smaller, puppy mortality increases, and dogs die before their time, but these problems are not generally viewed as symptomatic of a growing hole in the bottom of the boat. Instead, breeders diligently dig in to try to solve the problems, not realizing that most of what they are doing - more testing, ever more careful selection of breeding stock, keeping only the very best puppies to breed - is just making things worse. The thing that is important for breeders to realize is that you can't breed your way out of inadequate genetic diversity within a closed gene pool. Unfortunately, we need both education and leadership if we're going to turn this boat around. ICB is working on education, but I don't see any kennel club stepping up to the plate to coordinate what will need to be a coordinated, global effort to save these breeds.

Don't think there is trouble just in terriers. The inside poop on the other groups is coming soon...

So, you might ask, how did they decide that a breed with 100 registrations a year was at risk of extinction? They don't explain, but we can play with some numbers and shed some light on the question.

You might have heard of the "50/500 rule", which has for many years been the "rule of thumb" for population sizes in genetically managed populations. To establish a new population, it was thought that the minimum number of animals was 50, and to maintain a sustainably breeding population for 100 generations the minimum population size was estimated at 500. 

Here we need to explain the difference between census population size, which is the actual number of animals in the population, and "effective population size", which you can think of as the "genetic" population size. The 50/500 rule is referring to effective population size. What does this mean?

If the numbers of breeding males and females are the same, then effective population size will be the same as the census size. But if the ratio of males to females is not 1:1, then the effective size will be less than the census size. Let's say we have 100 unrelated animals, 50 males and 50 females. If each animal contributes one litter to the next generation and none breeds more than once, their individual genetic contributions will be roughly equal (allowing for some variation in litter) and the offspring will only be related to their litter mates. On the other hand, if we have 99 females and 1 male, again a population size of 100 animals, and each female produces one litter, they will all have to breed to the same male and the entire next generation will be half-sibs.

We can calculate the effective population sizes (Ne) of these two populations using a simple formula:

Ne = (4 x Nm x Nf) / (Nm + Nf)

where Nm and Nf are the number of males and females, respectively, bred per generation or per year.


So, for the first population, we have

Ne = (4 x 50 x 50) / (50 + 50), or Ne = 10,000 / 100; so Ne = 100

Effective population size is 100, equal to the actual number of animals in the population.


Now, for the second population:

Ne = (4 x 1 x 99) / (99+1), or Ne = 396 / 100; so Ne = 3.96

Even though the second population has the same number of animals as the first, the population "behaves" genetically like a population of only 4 animals with a sex ratio of 1:1. How can this be? The first generation will produce a population in which all of the dogs are either full sibs because they are littermates, or half-sibs because they share a sire. Those dogs will all have a COI of 0%, but for the next generation every dog will have to breed either to a full sib or a half sib. Their offspring will have COIs of either 12.5% (half-sib to half-sib) or 25% (full sib to full sib). In two generations, we have gone from 100 unrelated animals to a seriously inbred population. Now you can see the powerful effect of the popular sire

So you see how the 50/500 rule works, but here we need to interject and say that the rule of thumb that has been used by conservation geneticists for the last 40 years or so is now thought to be inadequate. Revised estimates are 100/1,000, or even 500/5,000, at least in part because survival for only 100 generations is not saving a population from extinction. In fact, we really should aim for sustainable breeding in perpetuity, and for this population sizes will need to be higher than the original estimate. The problem is that we can't really do the experiment that would test any of these predictions, so the best we can do is simulations of populations under various scenarios that resemble what might happen to a real population over hundreds or even thousands of generations. The minimum population sizes we should aim for to prevent extinction are the topic of hot debate, some of which you can read here.

While the argument rages, we can at least estimate some effective population sizes for the breeds on the Kennel Club's vulnerable list, just to get an idea of where they stand.

If we assume a litter size of 5, then 100 registrations per year would be about 20 litters. If we assume a different dam for each litter, that would be 20 bitches bred to 20 or fewer sires. We know we can estimate Ne if we know the number of males and females that produce litters from the equation as above:

Ne = (4 x Nm x Nf) / (Nm + Nf)

If the male to female ratio is 1:1 (i.e., 20 males and 20 females), then Ne will be 40. (Do the math and convince yourself that this is true.) This is far below the more recent recommendations of MVP size of 100. For breeds with smaller litters (e.g., 3), the estimate of Ne will be a bit higher, and for breeders with larger litters it will be smaller, but this is a useful ballpark, back-of-the-napkin estimate.

So for breeds like the Lancashire Heeler, Smooth Collie, Cardigan Corgi, and English Toy Terrier in the graphs above, we would expect Ne to be roughly 40. For the Skye, Sealyham, and Glen of Imaal terriers, with registration numbers of 20-50 per year, Ne will be 20 or less. For many of these breeds, the average litter size we're assuming (5) might be on the high size because of inbreeding depression, and if there are males siring more than one litter per year, the true effective population sizes could be in the single digits.

These breeds are indeed vulnerable, but this list only included the UK "native" breeds. How long would the list be if it included ALL the breeds with Ne less than 100? How long would the list be for breeds with Ne less than 1,000, which is thought to be the minimum necessary for long term sustainability?

The age-old maxim of "only breed the best to the best" takes on new meaning when you look at these figures. We are managing quality of offspring through selective breeding (both in terms of type and health), but we are not managing the size and health of the gene pool. To meet the minimum requirement of Ne = 100, we would need at least 50 bitches bred every year, and more than that if there were fewer than 50 sires. In fact, the maxim needs to be something like "breed at least 50 of the best females and 50 of the best males in each generation", and if we allow for some dogs siring multiple litters, it might be more reasonable as "breed at least 200 of the best females." You would still chose to breed the best to the best where you can, but you would also need to breed many dogs that are not "best" quality but just "good enough", and many of their offspring would need to go on and breed as well.

This would be a sea change in the way dogs have been bred for the last 100+ years, but it has to happen or many breeds are likely to collapse from inbreeding depression, excessive numbers of genetic disorders, and loss of alleles through genetic drift. These things are predictable from first principles of genetics, and decades of study of animal populations, both wild and domestic, support this scenario.

There is no chance at all that purebred dogs will beat the odds and persist despite closed stud books like the Chillingham cattle. The cattle managed to do everything right. For our dogs, we have been doing everything wrong. Dogs have not accumulated inbreeding slowly over many centuries as the cattle have, but very quickly over just a decade or two in most cases (see examples HERE and HERE). Dogs have not been bred randomly like the cattle, and in fact most individuals in a breed have not been bred at all. Unlike the cattle, the dogs with low fitness have not been culled from the population if they demonstrated other desirable traits. In breeding dogs, we have concentrated the deleterious alleles and selected against many animals that were genetically valuable but which didn't have just the right combination of alleles to qualify as one of the "best" that are kept for breeding. The Chillingham cattle are an extraordinary example of beating very long odds, not confirmation that we can continue to preferentially inbreed within a closed gene pool and expect to produce happy healthy puppies for generations to come.

Remember, the revised 50/500 rule is now felt to be too low, and newer recommendations are 100/1,000 or even 500/5,000, where the larger number is the size required for a sustainably breeding population that is expected to persist for centuries. For many dog breeds, these numbers will be impossible to meet, and for them we will need to adopt breeding strategies specifically designed to control inbreeding and minimize loss of alleles through genetic drift. Or we should consider the better alternative: convince kennel clubs to open the stud books so that dog breeding can return to the way it was practiced for thousands of years, with the occasional injection of new genes to reduce inbreeding depression and to improve the traits that need it.

How big is your breed?

-------

Errata: The original version of this article had an error, stating that a population of 99 females and 1 male had an effective population size of 40, which is incorrect. The correct effective population size would be 4.

-------

 Visscher PM, D Smith, SJG Hall, & JL Williams. 2001. A viable herd of genetically uniform cattle. Nature 409: 303 (doi:10.1038/35053160)

Hall SJG and JG Hall. 1988. Inbreeding and population dynamics of the Chillingham cattle (Bos taurus). J. Zoology 216: 479-493. (DOI: 10.1111/j.1469-7998.1988.tb02444.x)

1 April 2015
By Carol Beuchat PhD

After months of tireless negotiations, a group that includes representatives from dozens of kennel clubs all over the world, together with an expert panel of molecular, clinical, quantitative, and population geneticists, has agreed on a plan that will end the practice of the closed stud book, so that breeders can maintain the quality and health of purebred dogs with the introduction of new genes as they see fit. Registration procedures and issuance of pedigrees will remain unchanged with the exception that F1 and F2 offspring from all kennel clubs will be identified by a universal code appended to their registration number; special designation will not be necessary beyond the second generation. The F2 dogs will be able to compete and will be judged on their merits in all official kennel club events. In addition, a second code will be appended to all registration numbers that indicates the degree of inbreeding of the dog (calculated to 20 generations or as many as possible) by categories: A = 0-7%, B = 8-13%, C = 14-25%, D = 26-35%, F = 36-45%, and * = > 46%. These designations will allow breeders and pet buyers alike to judge the level of heterozygosity of the animal (including the immune system) as well as the potential risk of genetic disorders caused by recessive mutations.

These reforms will allow breeders to take appropriate steps to restore the genetic health of their breeds without the penalty of many generations of ineligibility for official kennel club events. It is expected that reducing the burden of genetic disorders in purebred dogs by allowing breeding strategies that have been standard practice for decades in the breeding of other domestic species will go far towards improving the image of the purebred dog in the eyes of the public, with the added benefit of alleviating the high cost of veterinary care that often falls on the owner of a beloved pet with an inherited disorder. Both breeders and kennel clubs are expected to benefit from renewed interest in purebred dogs by those who want to add a dog to their household that can be expected to live a long, happy, and healthy life. The group crafting the plan anticipates that this simple but long-overdue change will be a win for the kennel clubs, a win for the breeders, a win for the purebred dog owner, and certainly a win for purebred dogs.

May they live long and prosper.