Can Doggie DNA Tests Decode Your Mutt’s Makeup?

By Gemma Tarlach | November 25, 2014 9:00 am

Photo by Ernie Mastroianni/Discover

Mixed breed. Mongrel. Roadside setter. A something-something. Dogs of uncertain provenance get called a lot of things. When the animal arrives at a shelter, staff usually can make only an educated guess about the dog’s parentage.

Most of the dogs at my local animal control are assessed as “pit mixes” upon arrival — including the three I’ve adopted over the past 2 years. But a pit bull isn’t a breed: it’s just a type of dog characterized by a short coat, muscular frame and broad, oversized head.

All three of my dogs clearly — at least to my eyes — showed signs of specific breeds somewhere in their heritage: Tall and snow white Pullo looks like the breed standard for an American Bulldog. Tyche’s body is svelte like a boxer’s and inky black like some Labs. And lanky, long-limbed Waldo sometimes bays like a hound, especially when treeing squirrels.

Guessing my dogs’ breeds was a fun parlor game, but I wanted more definitive answers. So I turned to science. And, well, let’s just say it’s a good thing I didn’t place any bets on what was in my dogs’ family trees.


Finding Fido’s Family Tree

Consumer-targeted dog breed identification testing has been around for about a decade, with Wisdom Panel 2.0, owned by Mars Veterinary, as the dominant player on the American market.

Wisdom Panel looks at 321 genetic markers in your dog’s DNA to create a unique profile. That profile is fed into a program that assigns each ancestor from three generations — parent, grandparent and great-grandparent — into the best fit among more than 200 breeds in the Wisdom Panel database.

The doggie DNA test works differently than human ancestry tests, which typically trace both Y-chromosome DNA and maternally-inherited mitochondrial DNA back several generations. Instead, Wisdom Panel looks across the entire genome, not just on the sex chromosomes.

Wisdom Panel, like human ancestry tests, is more for the curious than for owners whose dogs are facing serious health problems. Medical DNA testing for dogs is much more focused and hunts for typically single-gene mutations that cause disease.

Urs Giger, a leading veterinary clinician and researcher, heads PennGen, at the University of Pennsylvania’s School of Veterinary Medicine in Philadelphia. He says that though Wisdom Panel is less focused, it still has some medical value. “Wisdom Panel’s purpose is to identify the breed or breed composite, which can be quite helpful information,” says Giger. Knowing your dog’s breeds can help your veterinarian tailor the treatment for a number of conditions, such as anemia, which may have different, breed-specific causes.


Photo by Ernie Mastroianni/Discover

On the Path to Wisdom

I read up on some of the research that led to testing for breed, like Wisdom Panel does, and wandered around the test’s consumer-friendly site, which is loaded with easily digested information, including the basics of the test explained in a video that even a kid could understand.

I decided I was ready to see what science had to say about my pups. I ordered the Wisdom Panel kits online for about $80 each (without mentioning my media affiliation).

DogDNA06Collecting each the DNA with swabs that resembled tiny toilet brushes — two for each dog — was a snap. Placing the swabs one at a time between cheek and gum, I counted to 20 and then popped them into the kit’s little holder to air-dry for a few minutes before sliding them into plastic sleeves to reduce risk of contamination.

I then activated each test online, a two-minute process that assigns a unique specimen number to each dog’s sample. I affixed the labels, closed up the kits and, the following morning, handed them over to my postal carrier. It was all easier than our average trip to the vet.

Who’s Your Daddy, Bitch?

Tyche’s report arrived first. Although she’s lean like a well-conditioned boxer, her head is classic American Staffordshire Terrier, the breed most commonly associated with “pit bull type” dogs. Perhaps not surprisingly, Wisdom Panel declared she was 50 percent American Staffordshire. Each of her parents was the offspring of a purebred AmStaffie, as fans call them, and a mysterious “mixed breed” parent.

I flipped a few pages past her shadowy family tree to a page called “Mixed Breed Signatures.” The five breeds Wisdom Panel identified as being most likely somewhere in the mix, so to speak, were, in descending order of probability, Standard Schnauzer, Curly-Coated Retriever, Great Dane, Irish Terrier and Saint Bernard.


Saint Bernard?

I saw nothing of any of those breeds in the black, blockheaded dog beside me on the couch, sniffing the report in hopes that it might be edible.

The next report to arrive belonged to Pullo, who everyone I knew had guessed was an American Bulldog. Bigger than an AmStaffie, with the snow white coat common to American Bulldogs, I was sure Pullo would be close to purebred.

And he was — just not purebred American Bulldog.

Wisdom Panel found he was 75 percent purebred AmStaffie — apparently, one great-grandparent ruined his otherwise snow-white ancestry. So…what was that bastard interloper?

Again I turned to the Mixed Breed Signatures page. While no other breed had crossed Wisdom Panel’s “detection threshold” — the line in the data that gives the lab enough confidence to include the dog as a definite ancestor — the breed closest was a Smooth Fox Terrier.


Photo by Ernie Mastroianni/Discover

According to the American Kennel Club breed standard, the Smooth Fox Terrier should not be more than 18 pounds. Pullo weighs in at 70.

The next four breed signatures potentially in his background were the Golden Retriever, Bull Terrier, Staffordshire Bull Terrier and Great Pyrenees, none of which I saw in the sleek-coated, big-boned doofus snoring in my lap.

Where in the World are Waldo’s Roots?

Then came the DNA detective work for Waldo, the most recent addition to the pack. I’d been fostering him over the summer, but no one wanted to adopt the big, dopey dog that I affectionately nicknamed Lurch. So I kept him.

Waldo has floppy ears shaped like perfect equilateral triangles and feet too big for his lanky body, which continues to grow. And grow. He topped 80 pounds a few days ago and shows no signs of slowing. Shelter staff told me they suspected he was part Great Dane. But his huge, baying bark and obsession with treeing squirrels suggested some kind of hound to me.

Eagerly, I downloaded his report.

Wisdom Panel found that Waldo was nearly half AmStaffie. The rest of his makeup was a mix of breeds (none of them hounds), but two signatures were strong enough to pass the detection threshold: boxer and Entlebucher Mountain Dog.


For anyone unfamiliar with an Entlebucher Mountain Dog (as if!), it’s the smallest of four regional herding dogs in the Swiss Alps, collectively called Sennenhunden. I looked up the breed standard and saw a compact, tri-colored dog with an all-business, very un-Lurch-like expression. It’s a rare breed, not exactly common to Milwaukee’s mean streets.

What was going on here? Was Wisdom Panel wrong?

I wanted answers, so I called Wisdom Panel’s veterinary geneticist, Angela Hughes.

Don’t Judge a Book (or a Dog) by its Cover

Hughes spends a lot of her time helping dog owners understand the test’s results, and is willing to admit the test sometimes barks up the wrong tree.

“I’m sorry, but that Entlebucher Mountain Dog is a false positive,” she said.

Having had a few days to ponder the reports by then, I felt a little disappointed. I’d already concocted a romantic, sepia-tinted backstory for Waldo’s great-grandparent, perhaps brought to Wisconsin by Swiss immigrants to wrangle a few of the plentiful ruminants of America’s Dairyland, then, one day, seduced by an urbanite boxer passing through town, running off to the bright lights and big city of Milwaukee.

“The Entlebucher is a rare breed with a strong signature. The computer could be fixating on one or two chromosomes,” Hughes explained, digging deeper into Waldo’s report, looking at breed signatures that popped up as possible ancestors but below the detection threshold. She decided that the Mastiff, suggested in his Mixed Breed Signature, was more likely to be in his background because it appeared on many of the lineage trees the computer builds as it processes a dog’s unique profile.

Hughes also had an explanation for why Tyche and Pullo’s results threw me for a loop. Wisdom Panel looks for genetic markers that don’t affect gene expression, she explained. The genetic markers aren’t dictating a dog’s behavior or appearance, and they’re not causing disease. Instead, they’re the product of random mutations that form a pattern over time in a single breed, just sitting quietly in the background of the dog genome.

By focusing on those “non-coding” markers, Wisdom Panel reduces the chances of a false positive — that’s how the company can claim an accuracy rate of 90 percent. It’s also how, Hughes says, owners can be surprised by their pets’ Wisdom Panel results.

“As humans we fixate on one trait,” said Hughes. “People think, for example, that bigger dogs with merle (a mottled coat pattern) are Catahoula, smaller merle dogs are Australian shepherds. But there’s a single gene for merle and it’s dominant — it could be from 20 generations ago, and everything from Schnauzers to Great Danes has it. We need to move beyond the single trait. It’s like seeing someone with blue eyes and saying ‘Oh, you must be from Sweden.’”



Everything Hughes said made sense to me as a science journalist, but as someone who has spent a lot of time peeling Waldo off trees in futile pursuit of squirrels, I clung to the idea of him having some large hound in his background.

Following up via email a few days later with Hughes, I asked for a better idea which specific marker might have caused all the Entlebucher confusion. Her reply answered my question by providing greater insight into the overall test:

The computer is making in the neighborhood of 7 million statistical calculations matching Waldo’s data to our database and what I see is a very distilled version of that analysis (otherwise it would take me days or months to wade through the data of a single dog).  So I don’t have that degree of granularity available to look at and be able to tell you that it is making a match say on chromosome A at position D.  What I can tell you is that some breeds (especially the rarer ones) have a genetic signature/pattern that is very “tight” – e.g. not a lot variation in certain areas – and so if a mixed breed dog happens to have a similar pattern or signature by chance (through genetic recombination and mutations), it can come up as a false positive match to that breed … I will tell you that the Entlebucher is not one I commonly see as a false positive (or a true positive for that matter!) so Waldo is definitely unique that way.

Oh, Waldo is unique, all right. Lurch is special, as are Tyche and Pullo, and any dog loved by its forever family. But, if the Wisdom Panel test told me only what I already knew for certain — that American Staffordshire terriers pop up in all three dogs’ background — was it a waste of money?

“I think it’s amazing that even the one breed was deciphered by 320 or so different markers,” says PennGen’s Giger, whose lab has no affiliation with Wisdom Panel. “The answer is so complex.”

Hughes notes that knowing the high percentage of AmStaffie in my dogs can also help me be vigilant about certain conditions, such as skin cancer, that they are more likely to get, regardless of what other breeds contributed to their DNA.

It’s also possible, note both Giger and Hughes, to run a much more precise breed identification DNA test, but it would cost thousands of dollars, well beyond what most dog owners would be willing or able to spend.

Giger theorizes that Wisdom Panel may be less helpful to owners of dogs — like pit bull types — that are the product of backyard breeders, who operate outside of AKC-registered bloodlines and often introduce other breeds to their stock in hopes of capturing a specific trait such as size, coat color or even regional aesthetics.

“One would have to moderate enthusiasm for the test results of a more locally generated dog,” Giger says tactfully.

I’m glad I gave Wisdom Panel a shot, even if my rambunctious but good-natured dogs are definitely of the locally-generated variety. It gave me a better understanding of the science that goes into breed identification — though I can’t help but note the shelter assessment (which Wisdom Panel says in its FAQ is generally only 25 percent accurate) nailed a three-out-of-three for my pups.

“In the big picture, people want a label,” says Karen Sparapani, executive director of Milwaukee Area Domestic Animal Control Commission (MADACC), where I adopted my dogs. “Some people were calling [the pit mixes] boxer-black lab mixes, or terrier mixes. But I decided it’s in the dog’s best interest to call it what it looks like. So, you know, if it looks like a duck, call it a duck.”

CATEGORIZED UNDER: Living World, Top Posts
MORE ABOUT: animals, genetics, pets
  • DJenny

    hahha interesting

  • Lisa Vale

    Strange that there’s no mention of either of the white dogs being Dogo Argentino as one clearly fits the profile, weight and markings. Maybe even both white dogs. Dogos are consistently misidentified at Pitbulls or AmStaffs, but they are a breed of their own.

    • Jennifer

      When I adopted my dog I was told he was a pit/hound mix, but did the Wisdom panel and he came back boxer/hound mix with a high percentage of Argentinian Dogo.

      • Ayla Falcone

        you were told that by who, the shelter? Shelters make anything up to but something on the “Breed” line.

        • Jennifer

          Yes, the shelter. I know they weren’t sure about it. They just guessed based his features and where he came from. Dropped at the shelter with his littermates.

    • Miss Cellany

      Maybe the DNA test doesn’t have Dogo Argentino in it’s database?

    • Miss Cellany

      Because Dogo Argentino probably isn’t in the wisdom panel database and thus can’t be detected.

      APBT is also not on the database so it’s actually not possible to test a purebred APBT and get a result that says 100% pitbull (I don’t know why people are testing their pit bulls and then being surprised the test doesn’t tell them “pit bull” – duh).

      APBT will normally come up as mostly AmStaff + mixture of other breeds (or unknown mixed breed) since there is greater genetic variation in APBT due to AmStaff founding stock being cherry picked from APBT stock when the breed started – a mix of random other breeds is how the test will interpret that extra variation.

  • Janice Koler-Matznick

    Since Mars will not test for purity of a purebred, pedigreed dog (only for parentage: matching to sire and dam), and because so many of their results come out with some extremely rare breed in the mix, like Irish water spaniel or Tibetan spaniel, I have no confidence in the Wisdom panel other than as entertainment. Many pure breeds have been derived from the same interbreeding landrace only recently (as in at most maybe 200 – 100 years) and so share ancestry. An example are the Tibetan origin breeds Lhasa Apso, Tibetan Terrier and Tibetan spaniel. Markers must be shared across such closely related and recently separated breeds, so why doesn’t Mars provide a list of all potential ancestral breeds rather than concentrate reporting the rarest? I doubt their results reach the 90% valid level. Last I heard (maybe 5 years ago) they used one litter of a known cross of two pure breeds to validate the test, and the panel did not even get all of them correct. I have asked the company many times for their validation test results/statistics, and they do not answer my inquiries, even when I tell them I am a professional dog writer. If someone knows where I can find more recent, complete information on Mars Wisdom Panel’s validity testing please post it.

    • Betsy

      I’m thinking of wasting some $ asking them to identify my purebred AKC Scottish Terrier–I’ll be interested in the results, eps. since he has championship lines on his sire’s side so there are papers back a ways on that side!

      • Jenna Farina

        Betsy that would be amazing! i always wonder if the breeds used to create the original “purebred” versions of dogs would show up in the dna panel too. i’d love to hear the results

  • handsome bob

    It’s worth clarifying that Pit Bull IS a breed, just not a breed recognized by the test, or some breed clubs–the American Pit Bull Terrier is a breed, and the only dog that should be colloquially referred to as a “pit bull.” Agreed that the longer-legged white dogs should have come back with Dogo, as the breed standard for APBTs or even Staffordshire Terriers dictates a shorter leg-to-body ratio. Agreed also that the prevalence of rare breeds and breeds not endemic to the geographic area of origin of a given dog casts serious doubts on the accuracy of Wisdom Panel’s tests. Interesting, additionally, that the black “Staffordshire” mix here presented standard Schnauzer in potentially detected breeds–my own nearly-purebred (or possibly actually purebred, since Wisdom Panel seems to have some hiccups) American Pit Bull Terrier came up with Standard Schnauzer as the second most likely result in the 5 potential breeds. The first? Boykin Spaniel. A breed detected so strongly that it nearly reached the detection threshold. Want to hazard a guess as to how many Boykin Spaniels are running around South Philadelphia? Oh, and then of course, Pekingese was third. Because people are always breeding Pekingese to their Pit Bulls. Now, I’m not saying that it’s impossible that those breeds are in there; goodness knows there are a fair number of intact, small, long-haired dogs roaming the streets, (probably not jumping fences and mating with Pit Bulls, though) but I am saying that I think some signatures mimic other signatures closely enough that ‘rare’ or unlikely breeds are reported far more often than they should be. You can probably count the number of Standard Schnauzers in Philadelphia on one hand, and the number of loose-running, unneutered Boykin Spaniels on one finger. I have another dog who I long felt was a Pit Bull-Boxer mix, as he exhibited many Boxer behavioral traits and aesthetic properties. His test came back reporting that he is nearly purebred AmStaff (in this case, I’ll allow that it may actually be AmStaff and not APBT) with the next top potential breed detected being Miniature Bull Terrier. He does have a bit of a bump on the bridge of his nose, giving him a peculiarly Roman-nosed look. So maybe Bull Terrier. But Miniature? The breed *after* that was greyhound, which is again possible… except for the fact that it’s rare enough that there’s an intact greyhound somewhere as their breeding is so controlled by the racing industry. Add the fact that he’s from Manhattan, and the number of seconds a greyhound would last on the street are in the single digits, and you may see the question. That being said, perhaps I am dubious because the next breeds were Papillon and Yorkshire Terrier. This is a 65lb dog–the tallest and longest dog that I have, in fact. The probability that he has either of those breeds in his heritage, or even miniature bull terrier is astoundingly low. Boxer did show up… as the last and least detectable of the 5 potential breeds.

    I think Wisdom Panel may want to address the high-incidence of reporting of rare breeds, and perhaps start just reporting the ‘groups’ of breeds that are detected, i.e. sighthounds, guarding, etc. It seems like that is what the test is picking up anyway, rather than the true likelihood of a breed.

    • Stephanie Eberhardt Wooley

      Pit Bull IS NOT a breed. It is a generalized term that encompasses many bully breeds. This is stated by ALL animal welfare organizations. Just bc APBT has “pit bull” in its name, that in no way makes them a Pit Bull. It makes them an APBT. The same goes for the AST & the ABT. This subject is highly debated among breeders and advocates. Breeders work hard on their bloodlines and should call their dogs by the appropriate breed name.

      • handsome bob

        Pit Bull is absolutely a breed so long as any breed organization chooses to recognize it and it is perpetuated separately from other breed identifications. IMO, it should not be a breed name as it’s a terrible name, but no one has yet come up with anything better and people are still fighting these dogs in pits in every seedy neighborhood in America. The importance of understanding that Pit Bulls ARE a breed is that it casts to light the misunderstanding that it is appropriate to call many different breeds of dogs ‘pit bulls’ for no other reason than that they all have short coats, blocky heads and bulldog heritage. If the American Pit Bull Terrier is not a breed, why on earth does it have literature dedicated to it, breeders breeding it, and dogs who cannot go by any other name identified as it? You can’t make a breed disappear just by saying it does not exist. APBTs are a distinct and recognized breed, and the reason that they are not recognized among many organizations is the stigma attached to the name, but the Staffordshire Bull Terrier which looks completely different from an APBT, is what the AKC and most DNA tests will recognize instead, which is ridiculous. My dogs are APBTs but identified on DNA tests are Staffordshire Bull Terriers because the DNA is indistinguishable, even if the morphological differences are dramatic. No educated person would ever mistake an English Staffordshire Bull Terrier for an American Staffordshire Bull Terrier, or an American Staffordshire Bull Terrier for an American Pit Bull Terrier. The primary difference that remains empirically between modern Staffordshire Bull Terriers and modern American Pit Bull Terriers is that APBTs are still popularly used in fighting pits and therefor have an addition layer of behavioral culling–human aggressive dogs continue to be culled, and only compliant dogs are bred. Staffordshire Bull Terriers, because they are recognized by large breed organizations, are bred more often as show dogs, etc and do not receive this street-level selection. If breeding for selected traits defines the way a breed develops (spoiler alert: it does) then the dogs being formed in the alleys and basements of America’s ghettos are a separate breed from the show dogs bearing a fancy British name that no longer applies to them.

        • Ayla Falcone

          If pit bulls (apbt) are their own breed why oh why can they be dual registered as both pit bulls and am staffs??? Not only can they be dual registered they can be shown as both! Since adopting a “pit bull” looking dog at a Philly shelter I’ve been trying to find an answer and failing miserably at it. If as you say their DNa is indistinguishable than wouldn’t that make them the same breed but maybe different type? We have multiple breeds with different types: show vs working spaniels, American vs English (lab, cocker). A working border collie looks nothing like a show bc. No one is arguing they are different breeds! From what I see it is the apbt enthusiasts that are trying to make the breed disappear by saying it is so rare and that all the am staffs are not apbts. I feel like the fact wisdom panel does not recognize them adds more fuel to their fire. I’ve witnessed owners saying “my dogs dna came out am staff so see he’s not a pit bull!” (nan a nan a boo boo). I’ve seen apbt enthusiast claiming unless you have actual paperwork explaining in detail your dogs pedigree from a reputable breeder you can not say hes a pit bull (even though he looks exactly like one). I mean really, would anyone argue that a dog that looks exactly like a lab is not a lab? The akc say they are different the ukc say they are the same wisdom panel says they don’t exist. I don’t think I’ll ever get an answer :/

          • Miss Cellany

            1) Amstaffs and American Pit Bull Terriers are the same breed. End of story.
            2) If you don’t have a pedigree for a dog but it looks like a purebred you can call it what you want but you won’t be able to register it with a Kennel Club so it makes no difference (this is true for ANY KC breed and even true for most working registries which will not let you register unless your dog has at least 1 working registered parent and even then would probably have to be ROM IF it passed a working test satisfactorily).

          • Darlene Molina

            I’m sorry, there’s an FCI recognized registry that recognizes the APBT as an actual breed? Don’t bother, the answer is no. Any group of people can call themselves a registry. Doesn’t make them anything more than a paper mill. If your dog is “registered” and you can’t transfer that registration to the registry of another country should you sell the dog, then your papers are worthless.

            People think Contenental Kennel Club papers are worth something but they’ll give you papers on a loaf of bread if you tell them it’s a Cocker Spaniel. Same with the multitude of other B.S. registries.

          • Ayla Falcone

            Umm.. first of all never heard of FCI, So I looked them up. Funny how they do not recognize the American Bulldog. Does that mean american bulldogs are not a breed? No. There is NOT ONE kennel club that recognizes every single breed of dog in the world. That is asinine. The AKC does not recognize many super rare dogs that are not popular in the US. Does that mean those super rare dogs that can only be found in Uzbekistan is not a breed. NO! Just because one kennel club does not recognize a breed, does not mean they do not exist.

  • gavin


  • toni haynes

    I just got a new 10 lb. fur baby from the local shelter in Jan. I’m thinking it might be fun to know her “pedigree”. I’m looking around for a test this week; in fact, I think my vet sells them.
    So far though, we see the long body of a dachsund, the soft curly grey coat of a poodle, the face of , well, maybe a yorky, but at LEAST some kind of terrier, especially around the eyes and muzzle….right now when people ask, we just say she’s a Sna-doodle…or is that a ta-doodle…or maybe a ya-doodle……I’m getting a headache now….lets just say she’s very cute, we love her oodles, and if you want a real one-of-a-kind dog, adopt from a shelter! Save a life!

    • Miss Cellany

      The funny thing is she may not actually be any of those things but may have inherited traits from other breeds that when combined caused her to look like those particular breeds you listed. For example there are many grey curly coated breeds (e.g bedlington terrier, Spanish water dog, bouvier des flandes, dandie dinmont terrier) and many long bodied short legged breeds (e.g. basset hound, dandie dinmont terrier, corgi, maneto podenco, even the extinct turnspit dog). She may even just be a Dandie Dinmont since that comes up as possibility for 3 of the traits you listed (curly grey coat, long bodied, terrier face).
      I see mongrels that look like border collies often but they don’t have the crouch / eye at all so probably don’t have any BC in them. Behaviour is often a good identifier of breed – but some individuals of a breed don’t display their breed behaviour so that isn’t a given either… Basically unless you have a pedigree for your dog you can never be certain what’s in the mix :)

    • 1calcium1

      Would love to see a pic of your dog because I just adopted an 8 lb dog from the shelter 2 months ago that I would guess are the exact breeds you mentioned but I have no idea! Waiting 2 weeks to get my results back. You don’t happen to be from FL are you? Maybe our dogs are siblings! LOL


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


See More

Collapse bottom bar