Understanding Ethnicity Estimates: How DNA Tests Calculate Your Percentages

An ethnicity estimate is one of the first things people see after a DNA test comes back. The chart that says “42% England and Northwestern Europe, 28% Scotland, 18% Germanic Europe, 12% Ireland” looks like a direct readout of ancestry. It is not. It is a statistical guess produced by comparing your DNA to a panel of people whose ancestry the company believes it already knows. Understanding how that comparison works changes how seriously you should take the second decimal place.

What an ethnicity estimate actually is

A consumer DNA test reads several hundred thousand single-nucleotide polymorphisms, or SNPs, from your autosomal DNA. The test then compares your pattern of SNPs to a reference panel, a curated set of people the company has identified as having four grandparents from a single region. By seeing which reference panel members your DNA most closely resembles across small windows of each chromosome, the algorithm assigns each window to a likely region of origin. The percentages you see are the share of those windows attributed to each region, rounded to add to 100%.

That is the entire trick. There is no library of “Irish DNA” anywhere. There are just present-day people with documented Irish grandparents, and your DNA either looks like theirs or it does not.

Why continental estimates are reliable and sub-regional ones are not

Two genomes from anywhere in Sub-Saharan Africa differ from two genomes from anywhere in East Asia in big, consistent ways. Continental-level estimates are almost always reliable for that reason. AncestryDNA, 23andMe, and MyHeritage all report continental breakdowns at high confidence, and the research literature backs that up.

Sub-regional estimates are a different problem. England, Scotland, Ireland, and Wales share enormous amounts of recent ancestry. Telling them apart inside a window of your genome requires reference panels large enough to have captured the small differences that exist, and statistical methods that can pull signal from a lot of noise. The companies do this with varying success. AncestryDNA reports the most granular British Isles breakdown, which is partly why it became the standard test for genealogy in the UK and US.

How the percentages add up to 100%

Every window of your genome has to be assigned somewhere, so the percentages are forced to add to 100%. That is a feature, not a measurement. If the algorithm cannot confidently call a window, it assigns it to the broadest-matching category rather than leaving a gap. A “100% European” breakdown is not a precision claim; it is the algorithm declining to break European ancestry into sub-regions where it lacks confidence.

This is why you can see “trace” amounts of a region you have no documented ancestry in. Trace estimates under about 2% are noise as often as they are signal. The reputable companies show confidence intervals if you click into the detail view. Use them.

Why people get different results from different companies

You can send the same saliva to AncestryDNA, 23andMe, and MyHeritage and get three meaningfully different breakdowns. The companies use different reference panels, different sized windows, different algorithms, and different region definitions. AncestryDNA’s “Scotland” and MyHeritage’s “Scottish” are not literally the same region; they are two companies’ attempts to define a category their reference panel can support. We cover the comparison in more detail in why ethnicity estimates change and ethnicity estimate accuracy.

How to read your own estimate

Look at the continental breakdown first. That is the part you can take at face value. Then look at sub-regional estimates with skepticism proportional to how much that region shares ancestry with its neighbors. British Isles sub-regions overlap heavily. Scandinavian sub-regions overlap heavily. West African sub-regions are coarser because the reference panels are smaller. None of this means your test is broken; it means the tool has limits, and the company that pretends otherwise is the one to distrust.

If you are using a DNA test to confirm a specific family story, the matching side of the result is usually more informative than the ethnicity estimate. The how DNA matching works explainer covers what the matches actually mean and how to use them, and the pillar guide to ancestry and genealogy ties the pieces together.