I keep seeing the statement in the title, usually in the context of a discussion about the need to improve genetic diversity of a breed with a cross-breeding program.

The people that say this reveal their poor understanding some basic principles of genetics that should be elementary level stuff for every dog breeder. But apparently not. This statement is false, and here's why.

Most of the hundreds of genetic disorders identified in dogs are caused by single, recessive mutations. A dog with one copy of the normal allele and one copy of the mutation will usually be unaffected and healthy. A dog that inherits two copies of the mutation will not, of course, have a copy of the normal allele, so whatever that gene is supposed to do in the body isn't going to happen. It will either be apparent as a disorder of some sort, or it will not be evident at all if the effects are subtle or do something like reduce fertility, or slow down some enzymatic reaction, or slow growth rate. But apparent or not, it can be expected that if a dog gets two copies of a mutation, there will be some sort of functional deficit.​
​

The fact that different breeds rarely share the same mutations is also the reason why mixed breed dogs are, on average, healthier than purebred dogs. While they might carry more mutations, those mutations are much less likely to be homozygous and therefore be expressed as disease (Donner et al 2018).

​The offspring of a cross breeding will produce offspring that will carry some of those mutations. If those dogs do lots of breeding, they will produce many copies of those mutations packaged in puppies that will enter the breeding population. The way to keep those mutations from being a problem, is to not make hundreds of copies and distribute them throughout the population. Keep them few and rare by nixing those popular sires.

REFERENCES

Donner J and others. 2018. Frequency and distribution of 152 disease variants in over 100,000 mixed breed and purebred dogs. PLoS Genetics  14(4): e1007361.  DOI: 10.1371/journal.pgen.1007361

​I asked the members of my ICB Breeding for the Future Facebook group to rate their understanding of genetic management and that of the people in their breed.

Like the folks that live in the mythical town of Lake Wobegon, where everyone is “above average”, most people responded to the "Rate Yourself" post with a score of 3 or more on a scale of 1 to 5. And pretty much everybody said that the overall level of understanding of the people in their breed was low (lots of 1s for this).
 
The latter response is very worrisome. Here’s why.
 
 Find your breed on this graph, which is the genomic (from DNA) inbreeding of a large number of purebred dog breeds. These data are consistent with data from several other studies of different populations, so we can assume that this is a fair representation of inbreeding in these breeds.
 
The green line is inbreeding of 6.25% (mating of first cousins), yellow is 12.5% (mating of half sibs), and red is 25% (full-sib mating). (I made this graph to highlight the data for Cavaliers for another post; the black line at about 41% is the average inbreeding of this breed.)
​

​This means that, because of the facts of basic genetics, you cannot “preserve” a breed in a closed gene pool. You might breed in a way that attempts to limit the damage to the gene pool, but you cannot preserve a breed if alleles necessary for function and health are lost from the gene pool every generation. You can be a “responsible” breeder," and make choices that attempt to limit the damage, but you cannot be a ”preservation” breeder, not when more genes necessary for the body to function are being lost every generation. 

Can we at least say that we are breeding responsibly, i.e., making breeding decisions that will limit the damage to a breed’s gene pool? Let’s look at some data.

These graphs are from Lewis et al (2015) and are based on the pedigree records of the UK Kennel Club. Note that the data were not digitized before 1980, so the graphs start there, and the COIs are much lower than actual values because the ancestors from 1980 back to founders are not included in the calculation. Also, ban on importing dogs into the UK was lifted in 2000, and the incomplete pedigree data for those dogs make it look like the average population inbreeding is going down after that, which is probably not the case.

The blue line is the average COI computed from the pedigree data. The red line is the level of inbreeding that would be expected if the dogs in the population were breeding randomly. If breeders were making a strong effort to avoid inbreeding, the blue line would be below the red line; if breeders are preferentially breeding dogs that are more closely related than average, the blue line would be above the red line. 

These graphs tell us about the overall breeding strategies being used in each breed. Breeders are preferentially inbreeding. (I have grabbed a few of the breeds that have a population large enough to show a trend instead of a line that goes all over the place.)

We want to preserve breeds. We want to produce healthy, long-lived dogs. But look at the data. How are we going to do this?

In response to the ban on breeding Cavaliers and Bulldogs in Norway, the clubs have responded that the breeders are in the best position to solve the problems and are working hard to do this. But inbreeding will continue because it is unavoidable in a closed gene pool. DNA testing can prevent disorders caused by a single, recessive mutation for which we have a test, but we do nothing to control the many other recessive mutations we know are lurking in the gene pool, and selection is doing a poor job of managing the problems that are likely polygenic. Plus, we are preferentially inbreeding, increasing the risk of producing genetic disorders from mutations we don't yet know about.

I noted up at the top that, in a survey of the people in my Facebook group, ICB Breeding for the Future, most people rated their understanding of the principles of genetic management and breeding for health at a 3 or higher on a scale of 1-5, where 5 is highest. They also overwhelmingly rated the average understanding of the people in their breed at 1. 

If we want to have purebred dogs, we need to solve some serious problems, we need to get it right the first time, and we need to do it soon. The kennel clubs say “We can do this,” but breeders assess the general level of genetic management expertise of their colleagues as rock bottom. 
 
I can probably count on one hand the number of dog breeders I know that have enough expertise in population genetics to “know what they don’t know,” and all of those people are professional scientists that happen to also breed dogs. Beyond that, there is an army of “Facebook experts” - people that might know something about little bits of a topic but dispense advice as if expert. Invariably, they say things that are incorrect, because they don’t know enough to appreciate the nuances, complications, and depth of concepts. They don’t know what they don’t know, and most people don't know the difference. These experts don’t do any topical coursework (e.g., I don't see them in ICBs online courses for breeders), and they usually have no background in science at all. What they “know” has mostly come from things they read on Facebook written by other people with no expertise. 

If you are tackling a very difficult problem, and if getting it wrong could result in catastrophe, these Facebook experts are extremely dangerous. Most breeders in my survey judged the expertise in their own breed as very poor. Most probably wouldn’t know good advice from bad; they are likely to be most swayed by things that sound “logical” or “make sense”. But their perspective is on managing the genetics of individual dogs. The genetics of populations are quite different, and indeed “the right thing to do” can often be counter-intuitive. You’ve heard that you should only “breed the best to the best”, but this will actually make it harder to improve traits and will ultimately lead to extinction of the population. In fact, this is why we are in this difficult spot, and continuing to use this breeding strategy is the hammer that will sink the last of the nails into the coffin. This is not an opinion; this is a necessary consequence of the mathematics behind population genetics. If you don't know this, then the "best to the best" advice seems like a good thing to do. But it's not.

The Norwegian kennel and breed clubs argue that they can fix the health problem of Cavaliers and Bulldogs by continuing to do the things they think will work. They haven’t worked so far, and they won’t. But apparently they don’t know that.

The Norwegian court offered that crossbreeding to solve the health issues would be allowed, but the feedback on Facebook has been adamantly opposed to even considering cross breeding programs. So, if the Norwegians have a plan to fix this without crossbreeding, I would like to hear how they will do it. 

The breeds in the spotlight in Norway have to come up with a plan to address the issues that put them in violation of the Norwegian Animal Welfare Act. The Norwegians are on the hot seat right now, but every breed has an inbreeding problem that is incompatible with sustainable breeding, incompatible with “preservation” breeding, and incompatible with health. What breeds have a plan to address their growing list of genetic health issues, which will only continue to grow? How will they know if their plan will work? What will they do if it doesn’t?

Purebred dog lovers face two huge challenges. First, we must fix the significant inbreeding problem that imperils  essentially every breed.

Then, once we have inbreeding down to a reasonable level, we need to breed sustainably, which means we have to  control inbreeding and loss of genetic diversity. To do this, breeders will need to understand population genetics, which provides the tools used for the genetic management of animal populations.

Are you thinking you already know a lot about population genetics? Among the essential topics you should be able to explain and discuss are these, for example: linkage disequilibrium, founder genome equivalents, effective population size, the Hardy-Weinberg equation, heritability, fitness, mean average kinship, observed and expected heterozygosity, and genetic drift.

One basic measure of the health of a breed is the size of the population. Large populations are likely to have more genetic variation than small ones (but not always!), and the large number of individuals buffers the population to rapid genetic changes.. This graph shows data for registrations by the UK Kennel Club from founding (1945) to the present. My data for AKC registrations are limited, but I have included them here for information.

The data for the UK Cavaliers reveal details about the history of the breed in that country. The population size increased after it was recognized by the Kennel Club, slowly at first, then very rapidly from about 1970 to 1989, when yearly registrations reached 15,833. From 1990, registrations have generally decreased through about 2010, and since then the decline has been steady and steep. In 2020, 2,967 dogs were registered, a level not seen in the breed since about 1970, a half century ago.

One concern about a declining population is loss of genetic diversity. This can be mitigated if breeders are careful to avoid inbreeding and they balance breeding across the population.  

The graph below displays the observed level of inbreeding in the UK breed population between 1980 and 2015 (in blue). The red, fuzzy line is the expected level of inbreeding in this population if the animals were breeding randomly over the same period of time. The pedigree database for this analysis (which is from Lewis et al 2015) only extends back to 1980, when digital records began. While this graph is useful to detect patterns in inbreeding levels in the breed population over time, the levels of inbreeding on the graph are much lower than what we know to be more accurate inbreeding from DNA analysis, which, as noted above, currently average about 41%

On this graph, if the observed level of inbreeding is higher than the inbreeding expected from random breeding, this indicates that breeders are preferentially inbreeding; i.e., choosing to pair dogs that are more closely related to each other than the average level of relatedness in the population. This also reveals that breeders could be producing puppies with lower levels of inbreeding by pairing individuals that are less closely related.

Note on the graph that the mean inbreeding coefficient decreases after 2000. This is the result of dogs imported after the restrictions to imports were lifted in the UK in 2000. These dogs have only three generations of ancestry in the UK pedigree database, so offspring they produce will appear to be much less inbred than they really are. Remember - in a closed population, the average level of inbreeding can ONLY increase over time. 

(These data are from Lewis et al 2015, and are slightly different from the registration records I have.) 
​

Inbreeding results from mating related dogs. Nothing can damage the genetic health of a population faster and more dramatically than a popular sire. This is a dog that produces a disproportionate number of offspring, populating the breed with a large number of offspring that all share half of his genes; i.e., they are all half-siblings. But, just like every other dog, the popular sire carries unknown mutations, and these will invariably result in a genetic defect a few generations down the road. Not all of these defects will be apparent to breeders; mutations that reduce fertility, shorten lifespan, or affect behavior are just a few examples of the ways these unidentified mutations can burden the health of the breed. 

​As in most breeds, the CKCS has popular sires that produce more than their fair share of offspring. These tables show that the most prolific sire in 2020 produced 30% more litters than the next most prolific sire, and 100% more than the dogs ranked 3 through 5. In terms of lifetime contribution, the top sire in 2020 produced twice as many litters as the second  ranked dog (111 vs 60). To understand the true consequences of popular sires on the genetics of a population see my article The Pox of Popular Sires. 

​

This chart also reveals the impact of the most popular sires in the breed. For instance, the top 5% of sires produce on average about 20% of the puppies in each year. Likewise, the top 50% of sires produce about 80% of the puppies, so the lower 50% of sires account for only 20% of the puppies.
​

Relating to the unbalanced use of sires is the tendency to restrict breeding of sires more than dams. This is revealed in this graph of the number of sires and dams producing the puppies of a particular year. In recent years, the number of females producing litters in a year has declined dramatically, which probably reflects at least in part the decline in the size of the population. The number of males producing litters has been relatively stable for most of the last 30 years, but in the last decade has been falling.
​

With the exception of the first graph, which includes data for AKC registrations for a few years, the data analyzed here are from the registry of the UK Kennel Club. However, this breed is popular in many countries in the world, and we cannot assess the true genetic status of the breed without including those data (or as much as possible) as well. If you have access to data for other countries, please contact me so I can incorporate them in these analyses. 

This is the first of what will be multiple installments that will shed some light on the genetic status of the Cavalier King Charles Spaniel. This information is essential to understand the history of the breed, and is even more critical or sound genetic management into the future. There are a few things already evident here that breeders should consider, perhaps most importantly, the declining registrations in the UK. As the country of origin, a healthy representation of the breed in the country has cultural importance. 

Related to questions about my previous Cavalier post, I still do not have the data necessary to determine the original size of the gene pool of the breed. I am chasing down ancestors and trying to connect ancestors, and I will let you know when have an estimate. 

Stay tuned - part 2 is coming soon!