Gear & Products
Dog DNA Tests: What Independent Testing Found
Updated July 30, 2026
By Morgan Reyes
The Short Answer
A journalist told a DNA testing company that a 52-kilogram purebred Great Dane was suspected to be a chihuahua. The test came back 100% chihuahua. Chihuahuas typically weigh under 2.7 kilograms. That's not a random error; it's the exact wrong breed the company was told to expect, which is a genuinely different and more troubling finding than ordinary inaccuracy.
| Test | What was submitted | What came back |
|---|---|---|
| CBC Marketplace, Quinn the Great Dane | Real DNA, suggested breed "chihuahua" | 100% chihuahua |
| University of Colorado, Lila the beagle | Real DNA, photo of a different breed | 50% poodle / 50% bichon (0% beagle) from one company; rottweiler from another |
The chihuahua test
CBC Marketplace built a deliberate test into its investigation of four dog DNA companies: Wisdom Panel, Embark, Accu-metrics, and DNA My Dog. They submitted samples from real dogs, and for one dog, a 52-kilogram purebred, show-winning Great Dane named Quinn, they told Accu-metrics and DNA My Dog that the suspected breed was chihuahua, a breed that typically weighs no more than 2.7 kilograms.
Accu-metrics returned a result of 100% chihuahua for Quinn, the exact breed Marketplace had suggested, not remotely close to her actual, confirmed breed.
Beyond that specific test, nearly every company gave different results for the same dog. Across the four companies and multiple dogs tested, results diverged case after case, despite every company separately marketing near-100% accuracy claims.
A separate study found the same pattern
A University of Colorado study, independently designed, found the same underlying problem: some companies appear to weight the submitted information more than the actual DNA. Researchers tested six direct-to-consumer companies using 12 purebred dogs, submitting cheek swabs alongside photos, with half the submissions paired with an accurate photo of the dog and half paired with a photo of a completely different breed.
A purebred beagle named Lila came back 50% poodle and 50% bichon frise, with 0% beagle, from one company, and as a rottweiler from another. The study's own conclusion: some companies may rely more on the submitted photograph than on the actual genetic analysis, which a real DNA test should never need in the first place, since genetics doesn't require a picture to read.
Why this matters beyond curiosity
A wrong breed-ID result isn't just a disappointing party fact. Breed-specific legislation is still in effect in many places, and the National Association of Insurance Commissioners confirms some insurers still use breed lists when underwriting homeowners and renters liability policies, despite acknowledging the studies behind those lists relied on visual breed identification rather than genetic confirmation. An unreliable DNA result carries real, practical consequences for housing and insurance eligibility on top of that, not just curiosity value.
Two separate investigations, designed independently, both found the same underlying problem: some dog DNA companies appear to weight what they're told to expect, a suggested breed or a submitted photo, over the actual genetic sample. A 52-kilogram Great Dane coming back "100% chihuahua," the exact wrong breed a journalist suggested, isn't ordinary test noise. It's evidence the test may not have been reading the DNA at all.
Bottom line: two independently-designed investigations of consumer dog DNA testing found the same pattern. CBC Marketplace's deliberate test, telling a company the wrong suspected breed for a purebred Great Dane, got that exact wrong breed back as a "100%" result. A University of Colorado study using photo manipulation found companies' results changed based on the submitted photo rather than staying anchored to the actual DNA. Given that breed-ID results can carry real stakes under breed-specific legislation and insurance policies, treating any single company's result as definitive is not well-supported by the evidence.
FAQ
Are dog DNA tests accurate?
Two independent investigations found real reason for doubt, including a company returning a result matching a deliberately wrong suggested breed rather than the dog's actual DNA. See the full test above.
Can two dog DNA tests give different results for the same dog?
Yes, routinely, per CBC Marketplace's four-company comparison.
Do dog DNA test results affect insurance or housing?
They can. Breed-specific legislation is still in effect in many places, and some insurers still underwrite using breed lists, per the National Association of Insurance Commissioners. See why this matters above.
Keep reading
A dedicated guide on evaluating breed-ID and health-screen test claims specifically is planned and not yet built.
A Note on Scope
This guide summarizes independent investigative findings on consumer dog DNA testing. It is not veterinary advice and does not recommend or discourage any specific product.
Sources
Image credits
- “Great Dane dog lying on grass in a sunny park with backpacks in the background.,” Matthias Zomer / Pexels, Pexels License
- CBC Marketplace, "How accurate are dog DNA tests? We unleash the truth". Fetched via browser-agent curl 30 July 2026 (direct WebFetch was blocked). Source for the Quinn/chihuahua result and the four-company comparison.
- Rando HM, Graim K, Hampikian G, Greene CS, "Many direct-to-consumer canine genetic tests can identify the breed of purebred dogs," Journal of the American Veterinary Medical Association, 262(5):1-8, 2024. DOI: 10.2460/javma.23.07.0372. Directly confirmed via Europe PMC (avmajournals.avma.org itself returns HTTP 403 to direct fetch). Source for the 12-purebred-dog, six-company design, the Lila/beagle result, and the photo-over-genetics conclusion. Authors are from the University of Colorado Anschutz Medical Campus, Smith College, University of Florida, and Boise State University.
- National Association of Insurance Commissioners, "Insurance Topics: Breed-Specific Legislation". Last updated 9/24/2025. Directly fetched. Source for the confirmation that some insurers still underwrite using breed lists, and that the underlying breed-aggression studies relied on visual identification rather than genetic confirmation.