How to Read DNA Test Results Without Misinterpreting Them

Your results are in. The dashboard shows percentages, a list of matches with strange names, possibly a haplogroup label, and (depending on the test) some health or trait reports. Most people scan it once, get a few “huh, really” moments, and close the tab. Reading it well takes a little more patience. Here is what each section actually means and where you should stop and not draw conclusions on your own.

The ethnicity breakdown

The pie chart is the headline of the dashboard. It is also the part most likely to be misread.

The percentages are a statistical comparison of your DNA to the company’s reference panel of people with documented ancestry from specific regions. They are estimates, not measurements. Two facts to keep in mind:

The numbers change. When AncestryDNA, 23andMe, or MyHeritage updates its reference panel, your percentages can shift by several points without your DNA changing at all. This is normal and reflects the panel improving, not the test being wrong.

The regions are broad. “Irish and Scottish” or “West African” are panel categories, not nationalities. Reference panels for some regions are much richer than others, which means precision varies. The same DNA at two different companies will give two different breakdowns, because they use different panels and algorithms.

What you can reasonably conclude: my DNA broadly resembles populations from these regions. What you cannot conclude: my specific tribe, ethnic group, nationality, or right to claim membership in a community is X. Identity runs through history, culture, law, and self-recognized communities, not a percentage on a website. We cover this point in more depth in our guide on DNA test accuracy.

DNA matches and centimorgans

Below the ethnicity chart is the matches list. Each match is another customer of the same service who shares enough DNA with you to be related.

The currency here is the centimorgan, abbreviated cM. It is an objective measurement of shared DNA segment lengths. Higher cM means closer relationship. A rough cheat sheet:

• 3,400+ cM: identical twin or yourself (a duplicate kit)
• 1,700 to 2,400 cM: parent or full sibling
• 1,200 to 1,500 cM: grandparent, aunt or uncle, half-sibling, double first cousin
• 575 to 1,330 cM: first cousin
• 215 to 650 cM: second cousin
• 45 to 215 cM: third cousin
• under 45 cM: more distant or possibly noise

These are ranges from peer-reviewed shared cM studies, most notably the work compiled by genealogist Blaine Bettinger. A single cM amount usually does not pin down a single relationship; it narrows the options to a small set. Treat the company’s “predicted relationship” as the most likely option in a list, not a definitive answer.

Matching is also database-bound. Your real third cousin who tested at a different company will not appear in your list. Uploading your raw DNA to MyHeritage or FamilyTreeDNA, which both accept free uploads, is the standard way to fish in additional databases. See our raw DNA data guide for how that works.

Haplogroups, if your test reports them

Some tests assign a Y-haplogroup (for men) or mtDNA haplogroup (for everyone) based on deep ancestral markers in either the Y chromosome or mitochondrial DNA. 23andMe, Living DNA, and FamilyTreeDNA report these to varying levels of detail; AncestryDNA does not.

For the broader buyer’s framework, see the guide to DNA testing pillar.

A haplogroup is a deep-ancestry label. It tells you which broad branch of the human family tree your strict paternal or maternal line falls on, typically reaching back tens of thousands of years. It does not tell you your nationality, your ethnicity in the modern sense, or anything about the other 98% of your ancestors. For a deeper look at the test types that produce these labels, see autosomal vs Y-DNA vs mtDNA.

Trait reports

Most tests include some trait reports: predicted hair color, eye color, freckling, cilantro taste, and so on. These are fun and largely harmless to read. Two caveats: trait predictions are probabilistic, and the predictions are usually less accurate for non-European-ancestry users because the underlying research was done predominantly in European-ancestry cohorts.

Health reports: the section where you should slow down

If your test includes health reports, this is the part of the dashboard where reading without context can lead to real harm.

A few principles, drawn from the consensus of the American College of Medical Genetics and Genomics (ACMG) and the National Society of Genetic Counselors (NSGC).

A consumer health report tests specific variants, not the whole gene. A “no variants detected” result for a condition like BRCA-related breast and ovarian cancer does not mean you do not carry a risk variant; it means you do not carry one of the small set the test checks. A clinical genetic test ordered through a doctor is a different category of product.

A “variant detected” result is not a diagnosis. It is one input. The clinical meaning depends on family history, other variants, and other risk factors that a consumer dashboard cannot weigh.

Do not act on a health result alone. Do not start, stop, or change medication. Do not schedule major procedures. Do not make decisions about having children based solely on a consumer carrier screen.

Do talk to a professional. If a health result concerns you, the NSGC genetic counselor directory lets you search for a board-certified counselor by location and specialty. Many take insurance. Your primary care physician is the other appropriate starting point.

Once you have read everything

Download your raw data. Every major service lets you, and keeping a local copy hedges against the company changing terms, going out of business, or shutting down a feature. Our raw DNA data guide walks through what the file is and what you can do with it.

Then, depending on what you found, work through our guide on what to do after DNA test results.