Leonberger history suggests that they should not be a genetically diverse breed. Although they were formed as a hodgepodge of large continental dogs a little over a century ago, two World Wars devastated their numbers and possible rescue outcrosses are mostly undocumented.
The breed is believed to be the creation of one man, Heinrich Essig, who was a dog dealer and traveling salesman; his claimed formula was Landseer Newfoundland x Saint Bernard, followed by more Saint Bernards and a Pyrenean Mountain Dog with the goal being a large white dog that was fashionable at the time. After Essig’s death in 1889 his nephew had the inspiration to promote the breed as a lion-like mascot for the town of Leonberg and established the final conformation of the dogs as having a lion-like rough coat with reddish-brown coloration, a black mask and black sable accents. Like most breeds, the origin mythology is poorly documented and stud books were not kept in earnest on the breed for several decades after its establishment.
As a German breed, the Leonberger was severely affected by both World Wars: only 5 known breeding dogs survived the first war and only 8 pedigreed dogs emerged from the second. Two severe bottlenecks like this are not conducive to preserve genetic diversity within a closed population.
That is why I was surprised to find that the 5 Leonbergers which were DNA tested as part of a study analyzing the genetic composition of Alaskan Sled Dogs showed that they had excess heterozyosity compared to Hardy-Weinberg equilibrium. Of 141 breeds tested over 96 genetic marker sites, there were only 9 purebred breeds found to be surplus heterozygous and most of those are only marginally so.
Note, this is not a measure of gross diversity in the breed, it’s a measure of which way the breed is being pushed genetically, either toward more conformity and homozygosity or toward more diversity and heterozygosity.
Genetic rescue requires a push toward heterozygosity before reaching a new equilibrium with the new alleles.
Inbred or Outcrossed? Bars that extend to the left indicate excess heterozygosity compared to the Hardy-Weinberg equilibrium (red line), suggesting active selection for allele diversity (i.e. outcrossing). The Leonberger is one of the most heterozygous breeds tested.
Here is the actual data represented in the chart. The purposely out-crossed hybrid sled dogs were found to be 20% more diverse than equilibrium and the Leonberger was second only to the Puli with over 10% excess heterozyosity.
| Breed | FIS |
| Sled Dog – Sprinter | -0.20197 |
| Puli | -0.11027 |
| Leonberger | -0.10662 |
| Cardigan Welsh Corgi | -0.05649 |
| Havanese | -0.04366 |
| Schnauzer Standard | -0.03506 |
| Norfolk Terrier | -0.0333 |
| Grand Basset Griffon Vendeen | -0.02908 |
| Collies (all) | -0.01916 |
| Dobermann Pinscher | -0.01482 |
I’ve only found one documented outcross on the Leonberger books: In 1954 a Leonberger stud Arko von Leonberg with a COI of 20% was bred to a Newfoundland dam Grisette von Bruckberg and one female from that litter, Alma von Rossbach, would cement herself into the Leonberger gene pool.
More recently European kennel clubs have registered Leonbergers with full breeding rights certificates “titre initial” (as opposed to a 3 generation provisional appendix registration “registre initial”) which are given after a number of criteria are met; typically an evaluation against the breed standard or significant show success, health testing, and perhaps even temperament evaluations. These dogs are believed to be pure-blooded Leonbergers instead of hybrid dogs which would normally take 3 generations to be admitted into the gene pool with full breeding rights and purebred status in registries that allow new blood. Sometimes they come with full pedigree and while the kennel club would void the known pedigree and perhaps calculate the COI of offspring as 0%, it’s not clear that either the registre initial or titre initial schemes are a source of new Leonberger blood.
Another possible source of Leonberger diversity lies in the history of Germany and the East after the wars. The Leonberger isn’t just a dog of the West and so it existed on both sides of the Iron Curtain. It is unknown and undocumented what sorts of possible crosses came into the breed during the Cold War before the reunification of Germany and the resumption of normal trade across Europe.
The last and least verifiable source of unexpected diversity would be intentional and unintentional pedigree fraud or error. It’s possible that some breeders when faced with a line that was not producing what they wanted or suffering from inbred disease or infertility decided to outcross, or an accidental litter proved sufficiently virtuous to register and the breeder either could not or did not choose to disclose this.
There’s also the possibility that I simply haven’t accounted for some factors which would have preserved heterozygosity in Leonbergers and that the bottlenecks did not have the expected effects on the gene pool. Certainly more questions than answers, but what an intriguing mystery it is.
* * *
Comments and disagreements are welcome, but be sure to read the Comment Policy. If this post made you think and you'd like to read more like it, consider a donation to my 4 Border Collies' Treat and Toy Fund. They'll be glad you did. You can subscribe to the feed or enter your e-mail in the field on the left to receive notice of new content. You can also like BorderWars on Facebook for more frequent musings and curiosities.
* * *
IIRC, the Puli is another breed that was down to very few and resurrected fairly recently. But the Doberman is interesting, given the widespred DCM. Also it appears that the sprinting sled dogs are diverse but NOT the distance dogs. I found my computer would not enlarge the original chart well enough to allow me to read all the labels, but it appears that these dogs are about the same as Belgian Tervuren (if TURV ? is tervuren)and less diverse than Malinois (BelM ?). I’m not sure if I spotted Groenendael (Bels????)but if so, the chart indicated they are less diverse by some than Tervuren — an indication of the problem about rules on interbreeding varieties (longhair to longhair and fawn to fawn but NOT longhair black to shorthair fawn) — to get to the genes in Malinois one has to go through some serious contortions if one has black dogs (Laken have it even worse). The new FCI rules may make life easier, but one still must get permission from both the country of origin (often possible) and the “receiving country” (in the US, a major issue).
I have glanced thru the article but haven’t read it intensively yet, so perhaps some of my questions as to why the difference between sprinters and distance sled dogs are answered there (my guess is there are more sprinting type dogs out there to use in a mix/match program). Peggy Richter
Pedigree-fraud is not unknown with some of these breeds which might explain the the skewed data.
I contacted the researchers and got the raw data. I’ll publish it with an upcoming post about the sled dogs.
it wouldn’t surprise me if caucasion owtcharka’s or blonde sarplaninacs or hovawarts were used inbetween the wars to save the breed. and maybe even in 1850.
it’s impossible to get solid coloured dogs out of 3 spotted breeds btw
The leonberger is one of the breeds the chinook people are thinking of bringing in to the chinook population which is too small. The chinook diversity genetic study is fascinating.
Again, cross breeding goes on all the time. I personally know in Vermont and QUebec that some backyard people are producing what they call alpine mastiffs, made up of leonbergers, newfs, anatolians, malamutes, and god knows what else. The new craze for designer dogs fuels this movement. Who knows if one of these ancestors came from these mixtures.
I’d like to know who the five collies were, or on what basis they selected the five dogs of each breed.
Not going to happen. I’ve already asked the researcher if there is any more granulated information on the individual dogs and that’s not available.
Where it says Collies it says “all” in parentheses. I could not read all the other breed acronyms. What is meant by “all?” All rough? All rough and smooth? Other collie breeds included or types, as in English Shepherd or old scotch? If no granulated info on individual dogs then perhaps this can not be answered? Not likely about the other breeds I mentioned I guess but any clues?
It probably means rough and smooth. I doubt they’d include English Shepherd or landrace dogs, since most people don’t consider them the same breed as collies.
Rough and Smooth are considered sepapate breeds in some countries, so that would clarify that they were being counted as the same breed. If the study were conducted in the UK, it probably would have separated the collies and combined the Belgians (separate breeds here, but varieties there).
One does wish that more than 5 dogs per breed were used. In Belgians, it certainly makes a difference if one selects from European dogs (where IV interbreeding is allowed) vice US (where you can import but IV interbreeding is not allowed) or even, within the US, if one uses animals from a given state (more likely to be related) or selected from across the country (less likely to be related). 5 Dobermans may not be statistically significant as a sample. Some of the earlier studies also used a fairly small sampling of animals but it’s pretty expensive to get a lot of samples for each breed. IIRC it was only when more samples were taken that the genetic input of Neanderthals was detected. It might be very interesting (but unlikely to occur) if some other group tried to duplicate this with different animals for each of the breed representatives. A replicated study would do a lot to substantiate the conclusions. The Leonberger results may remain unchanged (as might the Puli) given the initial small number and recent establishment of a registry, but some of the others might not.
“For each breed we utilized DNA samples from five individuals who were unrelated at the grandparent level.”
This could mean interesting things for the dogs that exhibit greater than expected diversity.
Jess recently posted..Goodnight, Sweetheart, It’s Time to Go
It wouldn’t be so difficul to find genetically diverse collies – there has not been much contact between UK, Continental, and US dogs. To add to that, in many countries there is no breeding between the rough and smooth dogs. Breeding a smooth to a rough would give as much diversity as breeding to a poodle.
Found this UK breeder’s website – she is one of the few that is using imported North American stock to lower her COI (this article states an average COI of 13.7% for collies in the UK). Her Canadian dog qualified for Crufts.
http://www.wicani.co.uk/breedingforafuture.htm
And the important thing (to the fancy, anyway) is that she’s getting dogs that compete and win in conformation. One dog who’s had a stellar career is Milas Collies’ Windhaven Stranger in Paradise. He has ancestors from Argentina; from dogs that were genetically separate from American dogs for generations.
http://www.milascollies.com/history.htm
Perhaps it’s not much of a cognitive leap to go from transatlantic outcrosses to crosses across breed lines!
In that first link, the breeder doesn’t understand what mtDNA is. Mitochondrial DNA is inherited from the mother. It’s true.
But it’s not magical DNA that is added to the 50/50 contribution from each parent. It is part of the 50/50 contribution.
It’s a very common misconception. You are not more closely related to your mother than your father because of mtDNA. It is just inherited only from the mother in the same way that one of your sex chromosomes is inherited via the father.
Retrieverman recently posted..Captive-reared New Zealand blue ducks released into a stoat-controlled region
I think the misconception about mtDNA came from assessing the puppies physically – maybe because they resembled the mother so strongly, it was easy to assume that she was prepotent in genotype as well as phenotype.
Yeah, but that doesn’t happen every time. Mendelian genetics shows that some traits are dominant, some are recessive, and you have incomplete penetrance with some alleles.
Some litters the pups will resemble the mother more. Others will resemble the father.
And others won’t look much like either.
I had a black dog that could have passed for Labrador cross, but her mother was an AKC registered golden retriever and her father a brindle boxer.
Retrieverman recently posted..Captive-reared New Zealand blue ducks released into a stoat-controlled region
If you look strictly at number of base pairs of genetic information, you might be able to distinguish a slight advantage to the mother over the father in male offspring. In humans (and most of Animalia really) the mitochondria have about 15k-17k base pairs and code for about 37 genes. The Y chromosome has 58 million base pairs and 70-200 genes. The X chromosome has 155 million base pairs and 900-1,400 genes. The whole human genome is 20-25,000 genes.
How we interpret this is actually very interesting. Males have less genetic information as they have one X chromosome and one Y and the Y is smaller. Plus they do get a few more genes from Mom than Dad via the mitochondria. But um, all genes are not created equal!
Men do get one X from their mothers, so they have one copy of those genes. But the smaller and funky Y chromosome has genes for male differentiation from the default-sex: female. SO, are men more like their mother than their father because they get more base pairs from them or are they more like their father because they get genes from dad which are active in specifically differentiating men in many ways from women?
As far as conformation dogs go, um I think this is a moot point because I don’t think mitochondria code for anything interesting in the show ring… look what they do, they code for the basic energy functions and metabolism within cells. Is this likely to shape the perfect head, to have a physical structure that is just so? I don’t think so.
Mitochondria likely play little to no role in sexual dimorphism because they are the same in both male and female offspring. They might have features which respond uniquely to hormones or something, but they probably aren’t the source of stuff showple care about in the breed ring.
Agree outcrossing is more than difficult in collies as one attempts to avoid the fad and fashion and favor of the day among show breeder champions. I have no doubt the chart is correct in collies, or darn close to it.
If one has spent any time doing extended pedigrees in collies the reality is right in front of you.
I hardly see the Argentina outcrosses as anything more than exports coming back as imports with Latin names as far as personal studies and in my opinion. Kathy Bittorf
Jana, I want you to understand that I’m not picking on you, but your comment has some common misunderstandings.
“It wouldn’t be so difficul to find genetically diverse collies – there has not been much contact between UK, Continental, and US dogs. To add to that, in many countries there is no breeding between the rough and smooth dogs. Breeding a smooth to a rough would give as much diversity as breeding to a poodle.”
All of this is entirely dependent on the founding population. How big is the founder population for collies? What are the founder contributions? How many lines died out? How many still exist? Were smooths separated out from roughs after the stud book closed, or do they have their own founder population?
If all rough collies in the world descend from the same founder population, you don’t get more diversity by importing dogs from other countries. That’s why it’s important to go all the way back in the pedigree. If smooths were separated out later, IOW, they trace back to the same founding population as roughs, then no, you will not get as much diversity by breeding to a smooth as doing a cross-breeding, even if smooths have been separate for many generations. What you may find, on analysis, is genes that may be rare in one population (roughs) as compared with another (smooths.)
There seems to be some misconception among breeders about COI and diversity. What lowering the COI does is slow the rate of gene loss. That’s all. It does not CREATE diversity. It preserves it. Once the stud book closes you have x amount of diversity, x number of available genetic combinations. At that point, you can only LOSE genes, you cannot MAKE MORE. A lower COI simply means you are losing genes at a slower rate.
Conservation breeding in endangered captive species is designed to preserve at least 90% of the existing genetic diversity in the population over a hundred year period. Dogs do not do this. Let’s take Afghan hounds as an example. The main founder population consists of 31 dogs. That sounds like a lot, and it would be, compared to many breeds, if each one of those founders had contributed equally to the subsequent generations. But they didn’t, due to popular sire syndrome, breed politics, fads, etc. Sirdar of Ghazni makes up approximately 30% of founder impact for most modern dogs. The existing diversity in the founder population was not preserved.
I could certainly import Afghans from the UK or Germany or Italy and they would not share common ancestors in the most recent generations of their pedigrees. But if I go back to the founders, they will share all the same founders as my domestic dogs.
“genetically separate from American dogs for generations” means NOTHING if the dogs go back to the same founders. Separate does not mean different.
Some really interesting misinterpretations in that first link but I will let Chris have his fun with that.
Jess recently posted..Goodnight, Sweetheart, It’s Time to Go
Jess, you are so correct. Those of us who have longer memories know how many of US collies were first purchased from Europe bloodlines. Then many exports to Japan, Germany,France, Canada, and yes Latin American countries.
Our Shetland Sheepdog friends have made several exposure from research that provided that all blue merle Shelties originated with cross breedings to Blue merle collies.
Extensive personal pedigree studies were so shocking ten years ago revealed to me a frightening need for some new “y’chromosomes. Kathy Bittorf
Kathy Bittorf
Oh no, pick away! I’m here to learn.
Even if Old Cockie may be in everyone’s pedigree, in the early 20th century one could enter a dog in Westminster with unknown parents. It’s not unthinkable that the random-source dogs would have been bred into the closed-studbook lines.
The origin of the smooth is kind of contentious – by some accounts, they derived from a different set of founder dogs. The rough-coated shepherd’s dog and the smooth-coated cur dog or ban dog. Thomas Bewick described them as being a separate breed: larger, stronger, and fiercer than the shepherd’s dog – better suited to working cattle and driving them to market.
Some books make it sound like the length of a puppy’s coat determined what his job would be (sheep or cattle), but what Bewick describes is more like two very distinct types within the landrace.
You are kind of missing my point.
It doesn’t matter if you could pick any dog off a farm and show it in 1879. If I look at the Crufts catalogs from the early 1900s, I can find lots of Afghan hounds and Salukis being exhibited. The point is that very few of these dogs made much of a contribution to the gene pool. Many were not bred, or their lines ended.
What matters is the number of dogs in the stud book when it closed. This is usually on acceptance of the breed by the registry. After this point, dogs must be registered to be included in the gene pool, and thus you can trace the founders, even if they were included (registered) after the stud book closed. A ‘founder’ is a dog that is ‘new’ to the gene pool, it has no pedigree behind it. We assume that these dogs are not related to the rest of the founders (which is not always true.) By doing a pedigree analysis, you can see the founder contribution. If one or a few dogs dominates that contribution, then you have lost much of the original diversity in the breed.
Even in a breed like Salukis, which have never really had a closed studbook due to desert bred dogs being including from time to time (and more of them, more recently), doing a pedigree analysis shows us that Western Salukis simply do not have a large gene pool, because breeders have followed the Tenents of Proper Dog Breeding, which guarantees gene loss:
Popular sires, dogs who sired a lot of puppies
Breeding very few dogs from each litter
Over-selection, applying very stringent criteria in selection of breeding stock
Line breeding for ‘type’
This is why it would be better for me to breed to a desert bred dog if I want ‘new’ genes, rather than importing a dog from Denmark which has the same ancestors as mine.
Jess recently posted..Goodnight, Sweetheart, It’s Time to Go
What about genetic drift? If you look at collies in the US, you’ll find the same ten or so popular sires over and over again. In the UK, there are also popular sires, but not the same ones.
I’m also saying that even with a closed stud book, it would be pretty easy to slip in an unregistered farm collie here and there.
That might account for the differences between UK and US collies – in the early days, the most “showy” show dogs were the British imports. The American ones were less “typey”, possibly from the influence of crosses with the working farm collies (the equivalent of desert-bred Salukis).
Closed stud books in both countries, but not a lot of dogs passing between the two. Salukis are a little different – it’s not native to the US or Denmark.
Genetic drift is just a change in the frequency of genes. It doesn’t make ‘new’ genes.
Collies are a numerous breed. Unless fraud in the studbook has been wholesale, a few crosses here and there aren’t going to make any difference, breed wide. After WWII, you had a major bottleneck in most breeds due to the death of the a lot of dogs; there was certainly a good deal of crossing going on, both above and below the board. There was also a new round of close breeding in order to set type again, so much of that new diversity was immediately lost again.
The only way to be sure of what you have is to do wholesale testing and find out. There is a commercial test called Optimal Selection from that checks, basically, for ‘alikeness’ on certain chromosomes. That’s about as close as you can get for dogs at this time.
You seem to be misunderstanding my point. If the dogs in the UK and the dogs in the US, for the most part, descend from the same group of founders, you may have differing gene loss in the different groups, but you will still have massive gene loss. Plus, within the show community, you have a lot of moving dogs around in more recent decades, from continent to continent.
What you would need to do, and actually, what should be done by every breed club in every country, is a deep pedigree analysis, going back as far as possible. I have one of these for Salukis, and one for Afghans, and you would think that the different countries would have very differing profiles in regards to ancestor contribution, but because all of these dogs start out with very small, closed populations, they don’t. With Afghans, the founder impact for the top founders varies by only a few percentage points, in the UK, the US, Sweden and Germany. Same with Salukis for the UK and US.
So you can see, that your few dogs being sneaked in would have very little impact, unless these dogs became popular sires and the population was still small enough for them to have a substantial impact. Once you get a large population, sneakage will have only a small local impact. Example: If I use an imported Saluki in my own kennel, I have incorporate a lot of new genes. But there are too many Salukis, worldwide, for this to have any real impact on the breed. It makes my dogs more diverse, but won’t disseminate out into the greater Saluki world.
Remember that Salukis have what amounts to an open studbook, so I know whereof I speak; imports, whether they be COO Salukis or farm collies, will have very little impact breedwide on a population that isn’t tiny. Like a kennel. The last import Salukis that had any founder impact were back in the fifties.
Appearance means very little. The genes that account for appearance are very few; I can change appearance in one generation. What we are interested in is the underlying genetic structure. Until a really good commercial test comes along, and I don’t know much about this Optimal Selection test, we are just guessing by looking at pedigrees and the few dogs that do get tested for the big studies. According to Christopher, who has read up that, the pedigree analysis almost always mirrors the genetic analysis, and often shows more diversity where there isn’t. Assuming, and counting on, some phantom dogs being bred in on the sly (or not) isn’t wise.
Jess recently posted..Goodnight, Sweetheart, It’s Time to Go
I want to make something clear I don’t think I did in the post. I don’t think that this table is measuring TOTAL heterozygosity or TOTAL diversity of alleles, rather it is measuring deviation from H-W Equilibrium. This is a rather unique presentation and I’m still working through all the implications myself and asking questions of the researchers to make sure I’m not butchering the presentation.
I know I haven’t done a post on H-W Equilibrium yet but let me give you an example of how two very different breeds can both be in equilibrium. Looking at one gene that has two possible alleles [A and a] (this can be extrapolated out for more complicated scenarios too), we can look at our population and take a survey of those alleles. We an get a gross count of how many of A and how many of a there are. We can also look at A and a in individuals and see how many AA, Aa, and aa we have. If we are in equilibrium, lots of random mating between our individuals will produce individuals with the same ratio of AA Aa and aa in the next generation.
There are lots of possible equilibrium ratios between a population that is 100% AA all the way to 100% aa. But realize that both saturated AA and aa populations are actually IN equilibrium.
So what this chart is telling is is not COI, not observed heterozygosity in the breed, but how these dogs compare to the equilibrium point that would be suggested by the gross numbers of A and a alleles. So it sort of speaks to where the dogs are being pushed genetically, not where they are on the great spectrum from homozygous with no allele choice to heterozygous with lots of alleles per gene.
So I think what this chart says is that most breeds are being pushed toward greater homozygosity, many in a big way. A few breeds are hovering around equilibrium, and only a few breeds are being pushed toward heterozygosity.
Collies, for example, are not starting from the same place as Border Collies. Using the data from a previous study (the one that established the 90-something markers) which I think is the same data used here for the non-Sled Dog breeds, the Collies show up as 38% expected heterozygosity. The Border Collies show up as 55% expected heterozygosity. Australian Shepherds show up as 61% expected heterozygosity.
Well, if we compare that with the chart in this post, the Collies are 2% on the excess heterozygosity side, Border Collies are 2% excess homozygosity, and Australian Shepherds are 18% excess homozygosity.
So what I think this tells us is that Border Collies and Collie are pretty stable near equilibrium, Collies are being pushed to be more diverse, Border Collies are being pushed to be more uniform, but Border Collies have a good degree more extant diversity of alleles than Collies do.
Australian Shepherds are being pushed hard toward conformity from a position of great heterozygosity.
Leonbergers were not included in the original study, but I’m trying to get the same comparable data from the researchers.
Well, it’s important to realize a few things about the Collie “diversity.”
One, it’s only barely on the heterozygous side of equilibrium. It’s safe to say that they would be better classified as “at equilibrium” than “excess heterozygous.” They are, however much closer to equilibrium than so many other breeds which are significantly homozygous and they are on the heterozygous side of that. It’s just not impressively so.
Two, they mushed together “Collie” “Collie Rough” and “Collie Smooth” into the category COLL. I don’t have more resolution to indicate if there were both rough and smooth dogs tested. It just might be that this recombination of dogs which are in some areas genetically distinct or isolated makes the recombined breed look more diverse. Easily possible.
Three, only 5 dogs. While this isn’t as limiting as one might expect for a low N, given that we are looking at 90+ genetic markers for each dog, it does make it possible that one outcrossed dog would bump up the numbers and since the way the dogs were tested does not allow for us to look at the individuals, we can’t really say “hey, that one dog is the source of all the diversity!” or “the diversity found was spread over all the dogs sampled.
Four, only 90-something sites tested. Newer and much more expensive testing might look at tens of thousands of sites in each animal. This might be worthwhile or it might not be. I’m not sure what the dog genome looks like overall, but it could be that these 90-something sites are a good indication of the diversity we want (they were selected by other research which found them significant in breed identification. One might find that huge chunks of all dogs’ genomes are homozygous and uninteresting (neither harmful nor unique between breeds… perhaps some of the most ancient and basic DNA that is present in all life) or not. I’ve never seen this sort of big picture stuff described in dogs and other animals. But, it does stand to question if these 90-something sites are meaty in looking at what we want to know. I’d say they are but it’s possible they are not. Even if they are not, diversity there might very well be indicative of diversity elsewhere too.
I agree Chris that the diversity differences in collies likely could be simply the differenes in rough and smooth. It is well recorded with Parader line that a smooth collie was purchased when Breeder could not find a rough for solution to a health issue. Thus bringing back once again the smooth collie here in the United States.
Well known breeding fact a smooth can produce a rough but a rough can not produce a smooth. Kathy Bittorf
Well, rough is recessive to smooth, so two roughs cannot produce a smooth, but a rough bred to a “pure for smooth” will produce all rough-factored smooths… 🙂
BTW, what about that claim that border collies are just as genetically diverse as “sled dogs”?
If you look at that chart, sprinting sled dogs are one of the most heterozygous types– much more so than the border collie.
Retrieverman recently posted..Captive-reared New Zealand blue ducks released into a stoat-controlled region
That’s a follow-up post (and my main interest in looking at this study in the first place), but it’s clear that Border Collies are bred like neither of the two sled dog strains. The one is clearly more inbred and the other clearly more outcrossed. This study is interesting because it shows that sled dogs aren’t even bred like sled dogs; there’s a clearly defined (in purpose and in genetics and in breeding styles) diversity within the breed and yet there is no push by that community to rename one of the types into a non-sled dog like there is by the sheeple to force the sport breeders to rename their Border Collies into something else.
There is also a time frame involved. If the dogs in a given breed in the US, UK and EU had been separated for say, 1000 years, one would expect some genetic drift by virtue of simple spontaneous / random mutations. The problem of course, is that this isn’t the case. Most breed have studbooks about 100 – 150 years old–not long enough for much genetic drift. If the great grandparents were the same, looking for dogs with different grandparents (3 generations) doesn’t do much. With only 5 dogs per breed, I’d be looking for “unrelated for 5 or more generations” if possible.
Wikapedia says “Leonbergers today can have their ancestry traced to eight dogs that survived World War II” — but did any breeding take place other than with those 8 dogs or were these all there were? It’s amazing that the Leonberger has a high diversity, but again, if the 8 dogs were bred to “unknowns” maybe it isn’t so surprising.
I’d love to see a study that looked for extremes within a breed –for example, are the field Setters and bench setters significantly different? I suspect they are. In many European breeds, numbers were severely reduced due to the wars and economic issues between and following them. Many of the registries “opened” to dogs that more or less met the standard but weren’t registered. In my own breed (Belgians), you can find “xxx” or “unknown” on post WW pedigrees. If these were truly dogs not much related to the registered ones, this infused a fair amount of diversity; if they were merely unregistered offspring of dogs in the registry, it didn’t. The new FCI rules MAY allow for more introduction of diversity or not — it will rather depend on how hard it will be to get the required permissions and how dedicated some of the breeders are to doing outcrosses (either IV or between “related breeds). Finding placements for the F1s that keep them intact and available for use in the registry is also an issue given some of the mandatory spay/neuter rules in the US, and certainly would require either money or other breeders cooperating with the project.
The time frame is interesting here. One, new genetic mutations are not the cause of this diversity, mutations are simply much too rare compared to how fast breeding schemes throw away diversity. It’s like the difference between a fire hose and a droplet from a sneeze. Only in the rare case where a new mutation is identified and valued and then highly bred on would we see a breed-wide increase in heterozygosity from it. The only example I can think of off hand is the new dominant semi-lethal tri-color mutation in “Panda” shepherds: it’s visible, novel, new, and people are probably trying to make a new breed out of it now.
Someone might try and make a breed out of that bully-whippet mutation for novelty. And there are perhaps other examples, but consider a small handful of examples over the millions and millions of dogs bred each year. Given all those millions of dogs, there are perhaps many new mutations, but only a fraction of those would even be noticed, few would be beneficial or desired.
I don’t believe that the history of the Leonberger would suggest that this is old heterozygosity that has been preserved over decades. I think this is much more indicative of very recent outcrossing, not deep diversity.
American hairless terrier. And that’s a recessive, so, increase in heterozygosity, no.
Jess recently posted..Goodnight, Sweetheart, It’s Time to Go
Well yes, but I was thinking of an example where the mutation was allowed to stay within the breed and thus actually add a small bit of new diversity. I think with the AHT, the mutation obviously got segregated out rather quickly which would mean that it adds no new diversity to the population where it started and then in the new “breed” made from the mutation, there is unlikely to be much diversity at all given just how fixated the breed definition is on one mutation.
Plus, you can’t get to 10% outcrossed based upon new mutations…. not a chance in hell that so many new mutations would arise over so many different loci.
The real answer here is that you are mixing breeds that have gone down a different path of throwing genetic diversity away.
That’s what I was curious about – how long would it take for genetic drift to make a difference?
I never thought that any “new” genes were created, but even though they started from common stock, there were very different forces influencing US and UK collies. Different popular sires, different bottlenecks, different fads. After all, the English shepherd, Border collie, sheltie and Aussie all started from very similar stock.
Belgian specific question: after the breed was decimated during WWII, how did they work the outcrosses? Did they use purebred dogs that were similar, like GSD? Dogs of unknown pedigree that resembled Belgians? Or was there a Continental landrace of herding dogs that they could dip into?
Jana,
If you’re talking about two populations that have the same founders that are split between nations with different breed standards, you are not going to get much genetic variation between the two as the result of genetic drift.
The two will be different because of founder effect and because of selection, but they won’t be dramatically different, unless one of them experiences outcrossing that the other population avoids.
About the only way new genetic material can come in one of these situation would be through mutation, and mutations don’t change that much of the DNA sequence.
It would take thousands of years before you had any kind of significant difference that resulted from each population receiving different mutations.
Retrieverman recently posted..Inbreeding and outcrossing in pigeons
BTW, I’m sure you’re aware that dog phenotype is almost always controlled by variation on just a few genes. American collies might not look exactly like British collies, but it’s likely that they are very closely related and there is probably very little genetic difference between the two.
However, there may some issues with rough and smooth collies in the UK and FCI. They are separate breeds in those countries, but they are separate varieties in North America. So you could have more diversity in rough collies in North America because they have smooths crossed in and vice versa.
But within the entire breed population as a whole, there won’t be that much variation.
Retrieverman recently posted..Inbreeding and outcrossing in pigeons
There’s also mathematics involved. It’s harder to have marked “excess” homozygosity when you’re already highly homozygous to begin with and have been for a long time. You might think of it like terminal velocity… after a while you adjust to the new normal and it takes a monumental effort to go even faster.
Collies are on the very inbred side of the scale and they are one of the first show breeds to have been fetishized, meaning that their refinement and inbreeding after popular sires occurred right at the start of the Victorian hay day for dog manipulation. It might be that they have simply become so inbred that much of their alleles are already at saturation level and thus their expected homozygosity is very close to the observed homozygosity.
The above chart doesn’t mean they are outcrossed and heterozygous. It means they’re very close to equilibrium.
Thanks. I think I came to the wrong conclusion from the chart.
The development of the show collie was so deliberate and so Victorian – it had to look just so, with these particular ears and this particular coat. And please, no merles, because only the common working dogs are merle!
Such a short period of history, and such vast changes brought to so many breeds. While the collie was “evolving”, the toy spaniels were crossed with pugs to produce shorter and shorter muzzles.
The chart above really needs to be paired with a measure of how inbred these dogs are to start with, which I’m TRYING to get, but so far unsuccessful in getting coverage for all the dogs. I have good data from the same lab on a subset of these breeds though.
When I get the data, I’ll do another post.
When you are starting from a point of limited diversity, you cannot ‘drift’ very far.
Genetically, within a breed you will have clusters, usually geographically, sometimes by variety when you have varieties (like Poodles and Doxies.) This does not mean that you have diversity between varieties, it means that you have certain alleles that are more common within the different clusters.
In the LUA Dal study that Mars did, Dals in the UK and the US cluster separately. This does not mean that Dals are diverse.
I was thinking in terms of how the St.John’s Dog was the root for both the labrador and the newfoundland – similar stock, but bred for different reasons and selected on the basis of different conformation.
Or is that not a good example, because of the crossbreeding that went on with the lab vs the newf, while collies have not had that input?
Greyhounds look pretty inbred — but I believe the authors didn’t include racing greyhounds, which would, I think, add diversity. What’s with great pyrs? If I read the chart correctly, they are right on the line. How did they manage to retain diversity? Ditto for the cardi corgi. I wish I could read the chart clearly. It looks like several breeds managed and one does wonder if there was a common factor there (temporary open registries, larger number of retained founders, selection criteria that is variable????? Why Cardis and not Pems???)
I don’t know if there’s much reason to believe that racing greys would be that diverse, they are chosen on a very particular metric, just like the distance sled dogs. Their short careers also suggest that they are not bred for long term health or hybrid vigor or anything like that. Certainly if you brought show and racing greys into the same breed as tested you would find something similar to what I suggest is the story with the Collies.
yes, I meant that if the study had included racing greyhounds and show greyhounds, they might show as “diverse” not because racing dogs are particularly diverse but because the two types have apparently selected for differing dogs for several generations — a 5 or 6 generation show pedigree probably wouldn’t have many (if any) dogs in common with one from a racing kennel. I think the same thing applies with “iron curtain” breeds — breeds like the GSD may have had entire lines segregated in West Europe vice those used behind the “iron curtain” — unfortunately as soon as the curtain came down, breeders have been busily integrating most of these dogs, but there might well be more diversity in these than some other breeds, even though specific lines might not have much. I note the Black Russian Terrier (I think I spotted them on this list) isn’t very diverse. Have you had any luck in getting the list in a more readable form?
It’s always possible that someone snuck in a dog that is different from what the pedigree states – AKC has announcements of those caught doing this regularly. But it’s also possible for “oops” events – some people really do think a bitch is secure when the dog is able to get to her via jumping 6’ fences, etc. I don’t think, however, that this small influx of new blood would be higher in one breed vice another in the long term. I think that if it wasn’t a case of differing lines/varieties (smooth/rough collies) being segregated for decades causing the apparent diversity then one must look at why there ARE some “closed studbook” breeds that seem to have managed a “zero” status or to have diversity. Clearly the distance sled racers are not particularly diverse even though they don’t have a closed studbook. And again, one might wish for more than 5 dogs or for a second study to try and replicate the results using different dogs (if the study is correct, one should be able to repeat it with the same results).
Jess said: “few dogs being sneaked in would have very little impact, unless these dogs became popular sires and the population was still small enough for them to have a substantial impact” –yes. it would be much easier for a few outside dogs to have an impact on Briards or Belgian shepherds than for the same # to impact German Shepherds. There aren’t that many Briards or Belgians compared to GSD or collies. On the other hand, finding a COO unrelated dog for Briards or Belgians is much harder (if even possible) than for desert sighthounds. Even where one has unregistered dogs (such as Altdeucher), care has to be taken that these sources aren’t eliminated by crossing in the other direction. There aren’t that many Altdeucher out there that one could be complacent about them remaining as a viable source of different genes. As for “new” genes, yes, mostly one does not find “new” genes. The hairless gene, the panda gene, these are “new” in terms of being new mutations not previously found. Mutation changes take place over far too long a time to provide much in terms of diversity. Inclusion of already existing genes (example being the Kbr “brindle” gene once found in Belgian Shepherds but selected against) and of course the other genes possessed by non related dogs is what is usually meant by talking about new genes in a given breed. It would be interesting to find out how related the Dutch Shepherds are compared to the Belgians — they certainly share some common foundation dogs, but how separated are they now?
5 Leonbergers is not enough to to get a statistically valid sample, though it is great that that small number of dogs has some heterogeneity.
The point in the article that I find exciting is that the Alaskan Sprint dogs were so genetically diverse, yet had a strong enough genetic signature to call a breed.
Perhaps this article should be taken as a lesson in a way to improve genetic diversity in other performance dogs.
Kate Williams recently posted..How did they breed dogs before the advent of the kennel club?
The Sprint dogs were very diverse. The distance dogs look to be about on a par with several “standard flavor” purebreds, despite the lack of constraints. I think that this may point out that while some characteristics may be easily found elsewhere, it may not be that others are. The sprint dogs used pointers as an outcross — fast, athletic, etc. But not so great in the environment — I note the use of “sweaters” for some. The distance dogs MUST be adapted to the environment plus have the other characteristics and this may mean that going “outside” is not so easy. One can do outcrosses that have F1 “not useful” and then regain use in F2 or F3 generations, but this does create an issue for a breeder in terms of maintaining the not so useful F1s (or even F2s) to get there. It may be that for the distance dogs, it just isn’t economically viable to do that.
Incidentally, I found it interesting that the sprint dogs primarily used ONE other breed for their outcrosses. The study indicated some others, but mostly pointers. That these dogs would have sufficent unique characteristics to be identifiable as a breed isn’t so surprising — the performance criteria does entail certain specific body type, trainabilty, coat, etc — which is, after all, why they use pointers and not just any fast athletic breed of the same general structure for their outcrosses.
Jana asked: Belgian specific question: after the breed was decimated during WWII, how did they work the outcrosses?
== officially, what they did was allow for breeding between varieties and temporarily recognized the results. (see http://www.belgiansworld.com/belgian-shepherd-history.htm). Farm dogs (what now would be called COO dogs) were allowed to be used. And probably there were some farm dogs that had Dutch, Altdeucher or GSD intermixed. The problem now is that these COO dogs have virtually disappeared. The Border Collies produced in the UK were cheaper to obtain than the native dogs and most farmers don’t really care WHAT dog they use — only a few are dedicated to a particular breed. Added to this is that while there are still some rural areas (the Ardennes is where the Battle of the Bulge was fought), and there are livestock still in Belgium, the number has been reduced just as in many areas. The dogs are pretty much used in protection work and more “urban” sports (agility, etc) although an interest in “hobby herding” has resulted in some trials here and there for those other than professionals. I’ve never seen any evidence that dogs from the US were ever employed (imports started very early in 1905, where the NY police department brought in a number of Belgians). Belgians were briefly popular in 1920s, got tarnished with being associated with police misuse of power and have never really recovered from this. Of course, popularity is a two edged sword, but it’s lack does mean a smaller gene pool than say, collies or GSDs.
Did they use purebred dogs that were similar, like GSD? Dogs of unknown pedigree that resembled Belgians? Or was there a Continental landrace of herding dogs that they could dip into?
Aren’t these results in statistical hot water due to probable sampling biases? Five dogs cannot represent a breed. Not related at the grandparent level is hardly a random sample. I’ve been trying to make sense of some Finnish studies of MHC/DLA diversity in different breeds (eg, http://www.suomenpartacolliet.info/DLA_raporttiEnglish.pdf) . . . they seem to use samples of 50+ dogs and I think they are able to select somewhat at random from canine DNA sample collections held at the University of Helsinki. Even then, they are careful to say their results speak only to the breed population within Finland.
Well, if you think of it this way, 5 dogs unrelated at the grandparent level could represent 20 different lines within the breed. We are also sampling just under 100 different sites for each of those dogs, so we are looking at a lot of alleles. 5 dogs doesn’t give a complete image of a breed, for sure, but that’s not really the goal here. You also have several factors about dogs that make 5 unrelated dogs more likely to be representative.
(1) Dogs are inbred, many breeds have homozygosity levels that are 60%+ meaning that 3 out of 5 of their genes are doubled up. That’s like a COI(full) = .60 … crazy levels of inbreeding. The more inbred a breed is, the more any one dog is representative of the breed as a whole.
(2) Breeds have genetic drift and selection against diversity. This means that for many genes there are very few alleles present in the breed. I think this study found on average only 2 alleles per locus! Like the first observation, this observation means that fewer dogs are likely to be representative of the breed as a whole.
Think about this at the logical extreme. If you had a breed that had no diversity, only 1 allele per locus, every dog was 100% homozygous, then one dog would BE the entire breed. (It’d have to be a male because we’d need both an X and Y chromosome). At the other logical extreme you wouldn’t have a breed, you’d have separate species and you probably wouldn’t have anything that looked like life as we know it because many of the most basic genes of life are shared across numerous species on the planet… so I don’t know if you could find two interfertile lifeforms that had no common DNA.
Anyway, I don’t think there’s anything wrong taking the data at face value for what it is, unless of course we have counter data, not just counter supposition. If we have a reason to believe that there is a particular sampling bias or systematic bias at work, by all means, let’s talk about it.
It’s impossible that a solid coloured breed stems from 3 spotted breeds. I guess the Hovawart or it’s ancestors played a big role too. And maybe the German Shephard dog in the time between the wars.
The Eastern European Leonbergers might have had an infusion of Sarplaninac or Ovtcharka to get better colours, more diversity, getting taller …
I’ve seen a pic or a Rottweiler x Leonberger that looks a lot like a Leonberger.
But then again, the Rottweiler is related to the Swiss Mountain Dogs, which are closely related to the Saint Bernard.