What DNA tests can—and can’t—tell you about equine ancestry - The Horse Owner's Resource

What DNA tests can—and can’t—tell you about equine ancestry

Experts explain how a DNA test could suggest surprising ancestry for a mustang from the American range.
Publish date:

Q: I have been following Dr. Deb Bennett’s “Conformation Insights” articles in EQUUS with great interest. I was motivated to learn my horse’s ancestry and sent samples of his DNA to Texas A&M’s Animal Genetics Laboratory for analysis.

The results came back with the first three breed markers being Selle Francais, Holsteiner and Hanoverian. This horse is a “mustang,” born wild in the Spring Creek Basin herd area
of the Disappointment Valley in Colorado. He was gathered and
adopted in 2011 as a yearling. My question is---how can this be possible? Is DNA testing really valid?

Lynda Larsen

Cortez, Colorado

Editor’s Note: Because of the many facets to this question, Deb Bennett, PhD, not only provided an answer but suggested that we enlist the assistance of other experts in the field of equine breeds and genetics. What follows are each of their perspectives on this question.

A:First, the results you obtained---no matter from how reputable a laboratory---could be bogus because of error. Samples can get mixed up or mislabeled. Even with the latest methods, I would not feel confident in DNA results returned from a single laboratory, but would submit to at least three reliable places and then compare all the results.

Second, the DNA analysis might be right on. Have you inquired as to whether anyone else has had a “mustang” from the same source herd and gathering tested? Many feral horses “born in the wild” in our country are not entirely of Iberian origin, and neither the Bureau of Land Management (BLM) nor any private agency is obligated to guarantee to adopters that they are “Spanish” to
any degree.

Much depends upon how many domestic horses of other breeds were turned out on the range on which your animal was born. It’s entirely possible that your animal was sired or mothered by a horse of non-Iberian origin. Many of horses on the range were simply unwanted or escaped domestic livestock. Whereas escapes date back to the early 1600s with Juan de Oñate’s colonization of Santa Fe, horses have continued to escape or be deliberately turned loose onto range at all times since. After the middle of the 19th century, most of those horses were not of Iberian origin.

Deb Bennett, PhD

Director, Equine Studies Institute

Livingston, California

A:This does get tricky. The basic “problem” is that the breed assignments are done by relative frequencies of specific allele combinations more than the absence or presence of specific individual alleles. Because most horse breeds share a great many of these, that means that few horse breeds have little bits of this information that are absolutely unique to them. The consequence of this is that some combinations that can pop up are more likely to be typical of some other breed than the one actually involved. 

So, while the technique works out well on average, it can be misleading for some very valid reasons related to procedure when used for a specific
few animals.

D. Phillip Sponenberg, DVM,

PhD, ACT (Honorary)

Professor, Pathology and Genetics

Virginia-Maryland College of

Veterinary Medicine

Blacksburg, Virginia

A: The DNA ancestry testing done by the Animal Genetics Laboratory at Texas A&M University is designed to give owners an idea of the breed or breeds might be in the background of a horse of unknown ancestry. The interpretation of results can be difficult and complex, but a pretty good explanation is provided on the facility’s website: vetmed.tamu.edu/animalgenetics.

The results from this horse’s ancestry test can easily be confusing ---or even make no sense to an owner
---but they are not unreasonable if you understand what the test result actually means. In this case the owner expects the horse to show up with “Spanish” ancestry because it is a mustang, but instead the test suggested that the horse’s ancestors were from European Warmblood breeds.

Both Drs. Bennett and Sponenberg gave reasonable explanations to the question “How could this be?” and “Is DNA testing really valid?” but I believe there’s even more to the answer.

First, although laboratory error is always a possibility because this work is done by humans, all laboratories that do equine DNA testing have very strict quality control and very seldom make errors. Also, sending a sample to three laboratories is overkill unless the first two come up with different types. Dr. Bennett does state that the analysis might be right on, and I strongly think that is correct and will explain in more detail below.

Dr. Sponenberg is also correct about how the assignments are made and that sometimes an individual may show closer affinity to a related breed than to the one it actually comes from. This is often an issue because an individual horse only represents part of the diversity of its breed, so that a statistical analysis of its genotype may place it closer to a related breed.

In this case the three breeds that were returned as the best fits to this mustang were European Warmblood breeds. All three have Thoroughbred ancestors. An analysis of the Spring Creek Basin herd that I did in 2007 showed that it was genetically most similar to what I call the Light Riding and Racing breeds, which include the Thoroughbred, Quarter Horse and all three of the Warmbloods that show up in this particular DNA ancestry analysis.

Of more than 200 wild-horse populations in the United States that I have tested over the years, only a handful show any genetic evidence of Spanish heritage. There is a suggestion that the Spring Creek Basin herd has some limited Spanish blood, but it is not likely to show up in the analysis of
a single individual.

The horse in this case is probably not directly related to any of the three Warmblood breeds but is a mixed blood animal that has some Thoroughbred ancestors, which is what the DNA analysis shows. Do read the explanation of results at the above website. The lab fully admits that this testing does not always give a reliable result but as this case shows, it usually does give a reasonable one.

E. Gus Cothran, PhD

Professor Emeritus

Texas A&M University

College Station, Texas