I don’t think UCD does breed testing. There is a private lab that does them and I believe a university in Texas.
1 Like
Now that you say that, it may have been in Texas. It has been 7-8 years now that I count.
I think it’s Texas A&M that does them.
1 Like
Texas A&M University calls it ‘ancestry testing’, not breed testing, and that seems to be a somewhat more accurate description. In my perception ‘ancestry’ is more broad than the specifics of ‘breed’.
The wall o’ words web page explains what to expect – but I wonder how many of the public understands it!
Excerpts … they start by explaining that from 400 to 600 horse breeds, internationally, have been identified by various scholars, and there are likely many more horses from somewhat isolated local sources, somewhere in the world, that could qualify as a ‘breed’.
So if you send in a sample from a ‘breed’ not in their list, you’ll get a result, but it won’t be very useful as they only compared it to breeds other than your horse.
From down in the word wall, highlights mine …
If a purebred horse is tested it will almost always be assigned to the correct breed. [Caveat - If it is one of the 50 breeds in their database.]
[If a mixed breed is tested … ] The results cannot give the proportion (percent) of the breed that the subject horse may have. That really isn’t possible because horses are so genetically similar. The test is reasonably good but there is no way to determine how accurate it is.
This is the thing to keep in mind against DNA testing of dogs … all horses are genetically more alike than are all dogs. Horse DNA genetically also works a bit differently than dog DNA (or rather, it might be more correct to say that the genetics of dog DNA has an inherent degree of flexibility that most other species don’t have).
Overall synopsis excerpts from the TAMU page …
… we selected 50 breeds that are most common for North America and also represent the major horse groups: draft horses; ponies; Oriental and Arabian breeds; Old World and New world Iberian breeds.
Chosen for being common in North America, and probably more distinct than many other breeds.
Horse ancestry testing at Texas A&M University is based upon comparing the DNA genotype of the subject horse to a reference panel of 50 horse breeds. Using a computer program based upon maximum likelihood analysis, the variants present at each genetic marker system tested for the subject horse are compared to those for each reference breed.
Also some North American breeds are not on the list, – example: Appaloosa, American Paint horse, because registries are open or partially open and allow crossbreeding.
Mustangs are also not on the breed list as it is now primarily a feral horse found in the western United States and managed by Bureau of Land Management (BLM). Originally mustangs were Spanish horses or their descendants, however throughout the years they had influence from many different horse breeds. There are several mustang registries, but overall there is just too much complexity to consider them in breed ancestry analysis.
The remark about mustangs reflects the general human attitude toward breeds (of any animal) – primarily, behind all the registries and all the selection of individuals, ‘it’s a breed because g-I say it is’. 
That is where all breeds begin.
Species have different, identiable, separate DNA traits. Breeds are a human construct, not a natural one.
For each breed comparison the probability that the subject horse came from that breed is calculated based upon the product of all the systems genotype probabilities. We then report the three breeds with the highest probability that the subject horse could have come from the breed in order of their probability of being an ancestral breed.
So that’s the caveat / buyer beware – it’s a probability test, not an exact test.
When a two breed cross is examined, the two parental breeds will almost certainly be given very high probabilities although not necessarily the 1st and 2nd assignments.
The more breeds involved in a cross the lower the probability that a good result will be delivered.
Also, understand that even though three breeds are reported that does not mean the subject horse has all three in its ancestry.
Another point is that breeds within a group of related breeds will be given similar probabilities. Thus, the subject horse may be half Belgian draft and half Suffolk but the test results may show Percheron or even a pony breed. That is because these draft breeds are very similar at the level we can test and the true pony breeds are closely related to the heavy draft breeds.
The remark about true pony breeds and heavy draft being closely related again comes back to the human conception of ‘what is a breed’. We tend to be focused on just the pony, or just a draft. And a bit blind to what is sometimes an ancient history of how the breed came to be and what types of horses were in the original gene pool.
1 Like
Exactly. There isn’t any one DNA profile = one particular breed.
Horse breeds can really be a mashup of different influences from past generations. People have bred horses by introducing a little of this, a lot of that, to get the results they are hoping for.
There is a whole other derail side discussion about open vs. closed books. A closed book that allows only registrations of foals from registered parents will tend to narrow down the DNA pool – maybe. IF it was a small pool to begin with. If the books has a large population of individuals, there will be far more DNA variations than a highy select, small population breed. That can go even further if the breed population is spread over a wide geographic area but tends to focus on breeding local individuals to local individuals. Although of course frozen semen and AI are diffusing that effect to a degree.
1 Like
All very interesting. The test I used was the Texas A&M one. Dogs were domesticated about 10,000 years before horses, so they’ve also had a lot more time to diverge genetically than horses.
All of this has got me thinking about breeding, not my area of expertise but its just so interesting to me that this horse who’s birth was probably an accident/not managed by people, is physically superior in almost every way over the thoroughbreds I’ve owned, whose pedigrees have been carefully managed for generations.
1 Like
TB are not bred for longevity and many sustain physical or emotional stress early in life. They are bred to be fast when young. I would say almost all drafty morgany stock type pony type horses that haven’t had something bad happen in early life are more sturdy than the average TB.
But they are not as fast.
I’d add that being semi feral is really helpful to long term soundness because he hasn’t been ridden into the ground already 
