The Science

How we read your dog's DNA.

Explore ancestry and trait predictions, genetic diversity and oral bacterial composition. Here's how to understand the results and their limitations.

What the report reads

Six sections, with context for each result.

The report combines sequencing evidence with reference-based predictions. Confidence varies by result, and some measurements may be unavailable. The current report does not include genetic disease-screening results.

The breeds behind the dog

Breed ancestry

The breeds behind the dog

Your dog's DNA is compared with a curated reference panel and reported as estimated breed percentages. Smaller contributions and closely related breeds can carry more uncertainty.

Coat colour

Colour and markings

Coat colour

Possible colour, pattern and markings where the evidence supports them. Some interpretations may use reported appearance; others cannot be determined. These limitations stay visible in the report.

Physical appearance

Coat and body features

Physical appearance

Predictions for coat and body features, each shown with its own confidence. These are genetic tendencies, not measurements of your dog.

Temperament and size predictions

Behaviour and build

Temperament and size predictions

Breed-informed behavioural tendencies and adult height and weight estimates. These do not measure your individual dog's personality or guarantee their adult size.

Genetic diversity

Inbreeding and reference comparison

Genetic diversity

A genotype-based inbreeding estimate, with a diversity comparison against dogs in the reference panel. It does not tell you how closely related your dog's parents were, and it is not a health score.

Oral microbiome

The submitted sample

Oral microbiome

Relative bacterial composition, community variety and monitored-bacteria comparisons. This is a snapshot of the sample, not a dental-disease screen or disease-risk score.

How breed detection works

Your dog's DNA, compared with reference data.

Published studies have documented genetic structure among sampled purebred dogs. In a controlled 2004 study of 85 breeds, microsatellite genotypes assigned 99% of the sampled dogs to their recorded breed (Parker et al., Science). That controlled purebred result is not a universal accuracy rate for mixed-breed consumer reports. Later genome-wide work (vonHoldt et al., Nature, 2010) examined dog population history, and in 2017 Parker and colleagues (Cell Reports) mapped relationships among 1,346 dogs from 161 breeds into 23 breed clades.

Consumer ancestry estimates compare selected genetic patterns with a provider's reference panel and model. Closely related breeds, small ancestry contributions, and backgrounds with limited reference data can be less certain. PuppyGenes reports probability-based estimates, not pedigree records.

Sample breed match

  • Labrador Retriever58%
  • Border Collie21%
  • Australian Shepherd12%
  • Chow Chow6%
  • Trace contributors3%

Illustrative. Your report shows your own dog's breakdown, dominant breed first.

Genetic diversity & inbreeding

How does your dog compare with the reference panel?

The report includes a genotype-based inbreeding estimate and a separate genetic-diversity percentile. The estimate depends on the genetic data and reference panel; it does not reconstruct a pedigree or tell us the exact relationship between your dog's parents.

  • A higher genetic-diversity percentile means a more diverse result compared with the reference dogs.
  • A lower inbreeding coefficient indicates greater diversity on this measure. The coefficient and diversity percentile are different numbers.
  • It's context for your dog's story and diversity, not a health score or a verdict, and never a diagnosis.

Illustrative diversity comparison

Less diverseMore diverse

Illustrative: more genetically diverse than about 62% of the reference dogs. The reference panel does not represent every dog population.

Lab methodology

From swab to report

  1. 1

    Cheek swab

    Collect saliva from your dog's cheek pouch for about 30 seconds. Skip food for 30 minutes and water for 10 minutes beforehand. No scraping and no blood draw are needed; follow the instructions supplied with your kit.

  2. 2

    DNA extraction and analysis

    DNA is extracted from the sample. Sequencing evidence is combined with canine reference data to estimate ancestry, included traits and genetic diversity. An inferred result is not a definitive direct reading of every genetic position.

  3. 3

    Analysis & interpretation

    Your dog's DNA is compared with breed reference data and analyzed for the included trait predictions and genetic diversity. Oral microbiome results describe the sampled microbial community. Each category is translated into a plain-language report with its limitations.

  4. 4

    Sample handling

    PuppyGenes uses a U.S. lab partner. Report data is stored in Canada and handled in line with PIPEDA. We don't promise destruction of the physical sample by a fixed date.

Where we stop

What our reports are, and what they aren't.

The report explains ancestry estimates, trait predictions, genetic diversity and the oral bacteria measured in the sample. Results can be uncertain or unavailable. These insights do not diagnose or rule out disease and do not replace veterinary care.

What we do

  • Review estimated breed ancestry
  • Explain coat and body predictions with their confidence
  • Explain genetic diversity and oral bacterial comparisons

What we don't

  • · Include genetic disease-screening results
  • · Diagnose or rule out disease
  • · Replace veterinary care
  • · Prescribe medication

The research behind it

Research behind the approach.

These papers provide background on canine genetics. Their study populations and findings are not PuppyGenes test specifications or an accuracy guarantee for your report.

Curious what your dog's DNA actually holds?

Explore ancestry, coat and appearance predictions, temperament and size estimates, genetic diversity and oral microbiome context. Results usually arrive within 4–6 weeks of posting your sample.