Here’s a new way of looking at dog breeds:
This comes from a newly released study that appears in PLoS Genetics.
More than 170,000 SNP’s were genotyped in 46 breeds and wolves. SNP’s are short for single-nucleotide polymorphisms, which is a DNA sequence variation that occurs when there is a difference in a single nucleotide between individuals in the same species. One can make comparisons about relationships through examining these variations.
The recent studies that found the red wolf to be mostly coyote and found that dogs were mostly derived from Middle Eastern wolves also examined SNP’s, but they examined only around 50,000 of them.
This one examined 170,000 SNP’s, so the resolution, so to speak, is much clearer.
Let’s look at the neighbor-joining tree. Here are some highlights that I think can be confirmed in the historical record.
- Modern flat-coated retrievers fit within the golden retriever breed. As I’ve noted, golden retrievers reflect much of the diversity in phenotype that existed within the old wavy and flat-coated retriever breed. Goldens and flat-coats share many foundational sires and dams, and one would expect that flat-coats would be closely related to goldens. The flat-coated retriever nearly became extinct in the Interwar Period, while the golden became extremely popular following the Second World War. This change in fortunes happened after the two breeds were officially placed in their own separate registries. Because the flat-coated retriever became rare, the modern representatives of that breed reflect only those flat-coats that survived that population crash. Goldens continued to reflect much more of the diversity in type and in genes that once made up the entire population– even though they are of a recessive color.
- Newfoundlands and retrievers are close relatives. I’ve always said that a Newfoundland is just a big retriever, but really, the Newfoundland is an offshoot of the St. John’s water dog that was imported to the United Kingdom from Newfoundland. It was selected to be a much larger dog than one normally found in Newfoundland, and it became a popular family pet for much of the nineteenth century. The golden, Labrador, and flat-coated retrievers derive St. John’s water dogs that were imported from Newfoundland and were selected to be gun dogs. Richard Wolters made the assessement that the large Newfoundland was derived from the St. John’s water dog, and this breed should be regarded as the common ancestor of the retriever-Newfoundland family.
- Nova Scotia duck-tolling retrievers are more closely related to border collies and Australian shepherds than retrievers. I have said that there is no evidence that they are related to golden retrievers, even though they superficially resemble dogs of that breed. I also suggested that we should think of the toller as a small retrieving collie. It also suggests that the recent cross-breeding of a toller to an Australian shepherd actually has greater merit than breeding it to another retriever.
- Dalmatians are related to Weimaraners, which are related to other pointing breeds. The postulate that the Dalmatian is derived from pointer crosses appears to have some merit.
- Border terriers are just one of the fancy varieties derived from the fell or Patterdale terrier. Patterdales are widely acknowledged to have Staffordshire bull terrier and bull terrier in them. But this map suggests that this infusion of bulldog or bull and terrier blood happened before the border terrier became fancy.
- Gordon setters are widely said to have border collie in them. It now appears that some Gordon setters do derive from border collies or from the ancestral collie that gave us the border collie, but others are more closely related to English.
- The theory that Large Münsterländers are derived from English or Gordon setters that were crossed with German long-hairs appears to have some merit. They are more closely related to one lineage of Gordon setter than to the Weimaraner.
- The pug may be more closely related to the spitz breeds because much of its development occurred in the West, particularly in the Netherlands. Spitz breeds were very common as pets in the seventeenth century Dutch Republic, and they were probably used to develop the pug in its present form. There were not likely vast numbers of pugs coming from China at this time, and the Dutch just used a local dog to make the breed more genetically diverse.
- The boxer is related to the bulldog, and we know that the bulldog was often bred into boxer lines. The boxer lines are longer because the boxer has had its full genome sequenced.
- This study does not necessarily show that German shepherds are derived from wolves. Rather the authors write:
The most obvious clustering of breeds is exhibited by two wolf hybrids: Sarloos and Czechoslovakian wolf dog, which exhibit a closer relationship to the wolf than other breeds as predicted by their known origin. The German shepherd also clusters with this group, although this is likely to be a result of its close relationship with the Czechoslovakian wolf dog, rather than with wolf. The tree is consistent with previous studies and supports the accuracy and reliability of the array.
Each time more SNP’s are analyzed, greater resolution is provided, and the result will be different. So let’s not assume that these results are going to remain unchanged.
However, I am amazed at how much this tree appears to confirm certain aspects that appear in various breed histories.
I tooooooooold you, you Duck Toller folks! They are collies!
And don’t say I don’t know crap about Tollers; they would had been my favourite dog if it wasn’t for the misadventure in Shibaness.
can you explain what the lines represent (color/length/size)? it’s a complete mystery to me
The length of the line just means that they have more of the genome sequenced. That’s why the wolves are so long, and the boxer, too.
They spelt Dalmatian wrong in the image lol.
I also wonder what the single green line inbetween the pug and JRT is for.
I’m still a little confused by the image (most likely due to sleep deprivation) so I’ve started reading about the study, but there is no way I’ll finish it tonight. =b
Maybe I’ll comment again once I’m done =D
It means there are some Border Terrier traits similar to Jack Russell Terriers.
Border terriers and JRT’s have been interbred quite a bit.
The generic working terrier in the UK is a Jack Russell with border terrier or something like it somewhere in its pedigree.
Here’s a typical commentation about how Dals “point”. It’s from the British breedclub discussion forum.
http://www.britishdalmatianclub.org.uk/forum/viewtopic.php?t=4147&view=next&sid=6f46fb44173692806fd094c9e352f6ec
I think Dals are rather mutts than pointers – mutts bred for the spots only. The reason they are linked with the Weimaren in the study may as well be that they both share the common history as an old Franko-Saxo hunting dog.
Many Weimarer still look like the old sturdy mastiff-type hunting dog that was used to run deers down.
DNA doesn’t lie.
Bridget, Bridget, darling. You need stop basing your theories based on phenotypic expressions. Dogs are extremely plastic in this regard.
And “mutt” shouldn’t be thrown around here either. Every breed started out as designer mutts. There is no such thing as a purebred dog which lasted for eons and eons forever and ever and ever. This is why naturalists donned the term “landrace”, in which cynologists hijacked to express their ideas where modern breeds and strains come from.
I can speak for my own breed here. A lot of Corgi breeders believe their dogs are either derived from dachshunds or the Vallhunds based on skeletons resembling teckels in Celtic graves; and dwarfed dogs found in graves in the Lapland. For the same reason, the Vallhund people believe their dogs are the ancestors of the Corgis because the Vikings did settle in Wales– nevermind the fact the they completely failed colonizing the area or the fact they were Hiberno-Norse; or the fact there are no Vallhund-like specimen in Ireland. History is useful, but the documents are limited in their interpretations– and unfortunately people give them too much credence.
However a recent study (Parker et al 2010) based on SNP sequencing on Alaskan Huskies showed Swedish Vallhunds are not related to the Corgis at all. In fact, Corgis are mostly terriers. In addition, the ULCA study in 2010 placed the Corgis with the Collie landrace by sequencing 48,000 SNPs. Furthermore, there was a study about a decade ago which analyzed the genetic markers in British and Nordic breeds and found the two are actually quite distinct regardless. Parker only drove the stake into the heart of the mythology several years later.
So, back to the Pointers:
Just because the Weimaraner have mastiff characteristics and there are historical evidences of Frank-Saxon dogs looking similar, it doesn’t means they are derived from the Roman mastiffs of yore. It makes sense why they genetically fall in line with other HPR breeds given their function. People breed like dogs with the same performance.
So, “Dalmatians” are not mutts. They are a breed, which falls specifically into the HPR group which Setters, Pointers, Munsterlanders and Weimaraners also fall into, bred specifically for its spots.
Keep in mind that, while it is true that Weimaraners didn’t start out as pointers, they were heavily interbred with them.
They started out as Thuringian bear hounds, but then the European brown bear went extinct.
So they bred them with the various HPR’s of central Europe to change them almost entirely.
It’s not to say you couldn’t hunt bears with one now. The Germans use their HPRs on all sorts of different game.
Davey-darling,
>>>>>>>>>And “mutt” shouldn’t be thrown around here either. Every breed started out as designer mutts. There is no such thing as a purebred dog which lasted for eons and eons forever and ever and ever>>>>>>>>>
– Yes, and in these terms, and also in Your terms as well, I see, the Dals of today are rather resent designer dogs – or mutts.
But have you looked at the table, dear Dave? What does it tell You?
If You look at the breed tree, the Weimer and Dal are tied together but distantly, for there is much more space between them than the space between the Lab and Newfoundlandin. So, do You think that the thing that once in the past was common to them – and still is, as the link shows – is the pointing ability? Or rather, floppy ears and body type or something like that?
I personally would have expected The Dal rather be next to the English Setter. But it’s not.
And finally see, how the Greyhound and also Irish Wolfhund are linked with the Retrievers – much more tightly than Dal with the pointing breeds.
The Dal / Pointer question – it really is so relative, and only relative, my friend. According to this study, I clearly won’t be tieing the Dalmation within the gun-dog breeds, like Retrieverman and You seem to do.
Davey-darling,
>>>>>>>>>And “mutt” shouldn’t be thrown around here either. Every breed started out as designer mutts. There is no such thing as a purebred dog which lasted for eons and eons forever and ever and ever>>>>>>>>>
– Yes, and in these terms Dals are quite designer dogs, or mutts.
Have you looked at the table, dear Dave? What does it tell You?
If You look at the breed tree, the Weimer and Dal are tied together but distantly, for there is relatively much more space between them than the Lab and Newfoundlandin. So, do You think that the thing that once in the past was common to them – and still is, as the link shows – is the pointing ability? Or rather, floppy ears and body type?
I personally would have expected The Dal rather be next to the English Setter. But it’s not.
And finally see, how the Greyhound and also Irish Wolfhund are linked with the Retrievers – much more tightly than Dal with the pointing breeds.
The Dal / Pointer question – it really is so relative, and only relative, my friend. Acording to this study, clearly, I won’t still be tieing the Dalmation with the gun-dog breeds.
Drop the phenotypic correlation. You can introduce the Corgi into the Boxer to get a bobtail, and get pure Boxer back in four generations. In fact, analysis confirming the purebred status of dogs are only as deep as four generations. Floppy ears are nothing special, nor is the pointing ability.
We are talking A-G-C-T nucleotides which most of the expressed bases are not even visible from the outside. Furthermore, there are limitations to the study as only 46 breeds employed in the studies, which is not even a tenth of all the recognized strains in the world. Note the English Pointer, Griffon, German Pointers are all missing. If other HPR breeds are also sequenced, it would paint a better evolutionary tree.
So where would you want to put them? With the wolves where the Sarloos or Vicak are located? It doesn’t make sense because the only breeds with noted wolf ancestry are derived from reindeer herding breeds [Finnish Spitz and Samoyeds]. The HPR group is a relatively new group of hunting dogs, so it makes more sense to put the Dalmatians with the HPR.
I should point out there are fewer strains of Retrievers than there are of HPR which is a glamorously diverse group.
I have no idea how to read this graph.
Why is the Toller fan so broad?
What is the Schnauzer doing so close to the Aussie?
What is the Border Terrier doing in between Boxers and Rottweilers?
I think each line represents a dog in the sample. Broader the line, the more dogs in the sample. (Dave may correct me, if I’m wrong.)
Schnauzers and pinschers were farm dogs– for my people anyway. They herded cows and sheep, as well as catching rats. They contributed to the Old German shepherd landrace, and Aussies are partly derived from German herding dogs, called “tiger dogs.” Tiger refers to the merle coloration, which is also what the Germans call dapple dachshunds– tiger dachshunds.
In the post, I make a suggestion that the border terrier, which is a fancy form of Patterdale-type terrier, also has the bulldog ancestry that Patterdales are supposed to have in them is not that recent. It was introduced at an earlier time.
There are old theories that fox terriers are derived from bulldogs that should at least be examined again.
I think it’s more like the bulldog has some terrier in it. Though I suppose it’s likely to have gone both ways.
I wish they’d sequenced the Staffordshire bull rather than the English bull terrier, I think it would be more revealing, since Hinks’ terrier was a deliberately created cross. Anyway, I’m still not totally clear on how to read the chart. Is it that the breeds are more closely related to breeds they are closer to on the chart? I’m a little intrigued at the Greyhound being next to the bull terrier. In my breed (Staffordshire bull terrier) I know someone who has a theory on Greyhound having been introduced into the early bull & terrier dogs. It’s certainly not impossible, though not currently possible to prove one way or the other.
This wouldn’t prove it either, since if there is a connection implied, it could have come about when Hinks was creating the EBT, he could have used some Greyhound.
Yes. The closer they are to each other, the closer they are to each other.
German tiger dog. Part of the Old German shepherd landrace.
More info: http://www.herdingontheweb.com/oldgermanshepherdinUS.htm
There is a merle dog in that list that looks like a blue merle GSD.
There is a organization in Germany for their old herding dogs:
http://www.a-a-h.de/?id=10
More photos here:
http://www.a-a-h.de/?id=21
The types of dogs they list are:
Fuchs
Gelbbacken
Kuhhunde
Schafpudel
Schwarzer Altdeutscher
Strobel
Stumper
Tiger
I have no knowledge in genetics (just enjoying this blog), but I thought this might be interesting – this Hungarian Mudi site describes history of some German, French, Hungarian, Croatian and Serbian herding breeds. http://www.netikka.net/laitinen/colors.html
Shepherds from these countries drove cattle and sheep all over Europe, so dogs have common history. Croatian Sheepdog looks exactly like Hungarian Mudi, but is a bit bigger and comes in black color only (my friend’s bitch is a rare black and tan). Some are born with short tails or bobtailed. You can get a Mudi puppy born in a Pumi litter.
A lot of these breeds today are still simple working farm dogs. Agility competitor here wanted to breed her Croatian Sheepdog bitch (amazing agility and Shutzhund dog) and searched for a male at farms all over Croatia – I find this quite charming. These are intense working dogs with big mouth!
Oh, Mudis and Pumis are already in AKC, Croatian Sheepdogs are only FCI recognized.
I just wrote something on the Harzer Fuchs subtype.
I think these dogs are pre-breed European dogs, and thus, they are worthy of genetic studies.
https://retrieverman.wordpress.com/2012/02/18/harzer-fuchs/
Pre-breed dogs from Southeast Asia and Africa have told us a lot about how dogs were domesticated, and because some of the oldest dog remains are in Central Europe, these pre-breed European dogs need to be examined.
https://retrieverman.wordpress.com/2012/01/31/african-village-dogs-point-to-middle-eastern-wolves-as-the-main-source-for-the-dogs-genetic-diversity/
Although the oldest dog remains are European, most European dogs are breed dogs, and they have lost a lot of diversity as they have become breeds. A pre-breed European dog might explain how dogs developed in Europe, and they might be related to these old European dogs that were around 14,000 to 30,000 years ago.
Thank you, I missed the post about the Fuchs!
The only rason I am familiar with Croatians and Mudis is that I am from Slovenia, bordering Croatia and Hungary and do agility. Croatians, Mudis and Pumis are hard workers with nice functional bodies and agility folk are discovering them. Croatian agility handler almost singled-handedly put Croatians back on the map in early 2000’s by running a great bitch. And then travelling through farms all over the country, helping people with breeding plans, searching for not related dogs. What I find so amazing is that many of these dogs belong to very simple farmers, but they all keep stud books (and are now FCI registered). This is a female helping with cattle, born with short tail:
Most have very curly hair, others are more flat coated. Some say Croatians have a bit of terrier in them, so a broader genetic study of early European dogs would be very interesting indeed. Luckily for Croatians, they are black (and loud) and in most cases go to working hands.
I’ve been told by Hungarians that Mudis are less healthy. In Mudis you can have all possible colors, merles and they breed to produce them (who wants the boring black only). Just wonder if color isn’t their peril.
Also, Pyrenean Shepherds (the small sized ones) are very popular with agility folk. A lot of breeders in the Pyreneans are breeding to dogs with no papers or have only 1 generation of ancestors in the papers (FCI). Genetics of these dogs are interesting too, due to geographic isolation (quite some merles there too).
Sorry to mess with your retriever blog:), but central European dogs are an interesting topic!
That’s okay. I’d say that more than half of my posts don’t even talk about retrievers. And a huge chunk don’t even talk about dogs!
I am of German descent, from territory along the Baltic Coast that once belonged to Germany but is now part of Poland. I have been interested in the German and other Central European herding dogs for a long time. There is some evidence that the tiger subtype of the Old German shepherd is behind the Australian shepherd, which, despite its name, was developed in the Western US. 20 percent of all Americans are of German ancestry. The next biggest ethnicity in the country are people of Mexican descent. There are more German Americans than English Americans. Our native herding and hunting dogs clearly have some German ancestry but how much is up to debate and conjecture.
This Tiger sure could be mistaken for an Aussie (sans big white collar and with a nice tail:)
Pyrenean Shepherds (Berger des Pyrénées) were used in Aussies too (French herders worked in USA West). Most likely Pyreneans and central European dogs come from the same stew.
There is also a Carea Leones, a Spanish “collie,” that could have played some role too.
http://carealeones.blogspot.com/2010/07/caracteristicas-fisicas-del-carea.html
Never heard of these!
What’s so interesting on the link, is those curly merle puppies. I’d say these are Mudis, no doubt (of course in reality Mudis are tiny dogs). And some of them could be Tigers. Others could be Pyreneans.
What a wealth of genes. Who would ever want a white Aussie!
Of course, that study has a low sample size of both schnauzers and Australian shepherds.
If they called them Austrian shepherds, they would be more accurate.
Many of the original dogs from the Old World were pinschers and collies, and when people migrate, they tend to bring their dogs with them. When they move, those dogs interact with other dogs in the area– and the New World is really a paradise for genetic admixture. You see this over and over again in colonial Latin America, and post-colonial United States where citizens claim not only one ethnic identity but take pride in derived from multicultural background.
Seeing how the Schnauzer is a wire-haired Austrian or German Pinschers, it makes sense why they are partially related to Australian Shepherds.
Such diagrams are interesting (as far as they go), but any two-dimensional graph seems far too simplified. Actual relationships among breeds are undoubtedly much more complex in many regards, but there would be no easy way to depict all the information graphically.
And what might be represented by that unlabelled yellow lilne between the Flat-Coated Retriever and the Irish Wolfhound?
I don’t see this line, but there is a golden retriever on the other side of the flat-coated retrievers.
This isn’t necessarily a phylogenetic tree. You’d need more breeds to draw it properly.
>>>>>>>>Actual relationships among breeds are undoubtedly much more complex in many regards, but there would be no easy way to depict all the information graphically>>>>>>>>>
Dear M.R.S., I defenitely agree with You. Every new study and graphics draws a little different picture. Think it’s not only because more samples info are processed, but also because of the quality of that. In this study they concentrated partly in the physical trates and the genes that produce them – so that naturally shows in the results.
When Retrieverman says >>>>>>>>You’d need more breeds to draw it properly.>>>>>>>>
Does he mean “all the breeds, all the genes and their relations”? Well I’d say it’d be impossible to draw a diagram about it if not to compare that all in one time!
Don’t you guys think this one either shows the whole story at all. It shows those agents the researchers have chosen to be studied. Also, it clearly reflects the fact that behind the dog genome studying boom is the need to search for some answers for the human diseases.
You need more breeds than 46. There are certain groups that are under-represented here. There are only two sight hounds, and both of these are from the British Isles.
[…] here is a fascinating new way at looking at the relationship between different breeds of dog. You can find it on the Retrieverman’s blog. It may surprise […]
Tervs are related to Beagles?????? Uh, don’t think so. I can possibly buy off on the Toller being decended from a common ancestor but previous studies put the Belgians as closer to the Irish Wolfhound. Or am I reading this thing wrong? While Belgians were part of the study, the text pretty much ignores them (as usual). Peggy
The study also contradicts OTHER studies using DNA that found different relationships between Belgian shepherds and other breeds. So DNA may not “lie” but it certainly does matter what genes one looks at. It’s hard for me to come up with a breed that Tervs (not Belgian Shepherds, as the study apparently only looked at the fawn longhairs and not the other variations on the theme) resemble less than a Beagle. The structure is different, the temperament, the function and even the location of origin. I’d love to see a study involving Briards, Beacerons, Dobermans, Picardy shepherds, Mudi, Puli, Belgians, Bouviers (we know the Belgians & Bouviers had some common foundation dogs) des Ardennes and des Flandres and the Schnauzers with perhaps GSDs and BCs as the outgroups. But I am either reading this article wrong or they got something wrong because I just do not see any relationship between Belgians and Beagles other than their names start with the same letter.
Peggy.
[…] One can also see how dogs like these would have played a role in developing the Gordon setter, which is well-known to have some collie blood– and this fact has been confirmed in a recent genomic study. […]
[…] There are no proven facts that Chihuahuas are derived from anything other than wolves. All the evidence shows they are toy dogs, and at least with AKC Chihuahuas, they are primarily derived from European dogs. […]
[…] study, which unfortunately didn’t include Brussels griffons, pekingeses or shih-tzus, found a close affinity between the pug and the Jack Russell terrier. This study examined an even broader sample of the genome than the earlier 48,000 SNP study. It […]
[…] I don’t have much evidence for this theory, other than the genetic studies on Dalmatians have nested them within the pointing gun dog breeds. […]
[…] New genome-wide study sheds light on dog breed relationships … […]