Raw DNA Data: What It Is, How to Download It and What You Can Do With It

What a raw DNA file is

A consumer DNA test does not read your whole genome. The lab uses a genotyping chip that checks a fixed list of positions, typically between 600,000 and 700,000 single-letter variations called SNPs (single nucleotide polymorphisms, pronounced "snips"). The raw DNA file is the plain list of those positions and the two letters (alleles) you carry at each one. There is no interpretation in it: no ancestry percentages, no health report, only data.

Each line is one SNP: its label (an "rs" number), the chromosome, the position on that chromosome and your genotype. An illustrative line looks like this:

The file is yours to keep and to reuse. Every company's own reports are only one reading of it, and other services can read the same file, which is why you usually do not need to take a second test to get a different kind of analysis.

How to download it from each provider

Every major provider lets you download the raw file for free, but the menus change from time to time. If you cannot find the option, search the provider's help pages for "download raw data".

File formats compared

The same information is stored slightly differently depending on the provider. This is why a file made for one service sometimes fails in another, and why a good reader detects the format instead of asking you to convert anything.

The sizes are approximate and vary by chip version. Older kits and newer kits from the same company can differ.

What it can and cannot tell you

What it is good for

• Common, well-studied variants. Many of the variants that affect how people handle nutrients, caffeine or lactose are common, and a chip reads them reliably. Examples are MTHFR, CYP1A2 and the lactase region.
• Tendencies, not verdicts. A variant shifts the odds or the way you process something. It rarely decides an outcome, and for nutrition the effect of any single variant is usually modest compared with what you eat, how you sleep and how you move.
• Some pharmacogenomic markers that can be worth discussing with a doctor or pharmacist, never a reason to change a medication on your own.

What it cannot do

• It is not your genome. A chip reads a small fraction of the variants that exist. A "not found" or "normal" result on a gene does not mean the whole gene is normal, because rare variants are mostly not on the chip.
• Rare variants need confirmation. A 2018 study published in Genetics in Medicine found that about 40% of rare variants reported in consumer raw data were not confirmed by clinical-grade testing. Common variants do not have this problem, but anything that looks alarming on a rare variant should be checked by a clinical laboratory.
• It is not a diagnosis. If you have symptoms or a family history, the route is your doctor and a clinical test, not a consumer chip.

Why two files cover different genes

Each company chose its own list of positions, and each has changed that list over the years. So a gene that one file covers well may be only partly present in another. This is a property of the chip, not a fault in your file.

We measured it on our own panel. Out of the 118 variants shown in our reports, an AncestryDNA file we used as a reference contained 94, while a Family Tree DNA Family Finder file contained 53. When a variant is missing from your file, our reports mark that gene as not covered by your chip instead of guessing. If you have several kits, the one with the best coverage of the genes you care about gives the fullest picture.

Privacy: what to check before you upload

A raw DNA file is genetic data, a special category of personal data under European law, and it is unique to you and, in part, to your relatives. Before you give it to any service, check these five things:

1. Where it goes. Is it processed on your device, or sent to a server?
2. How long it is kept. Deleted after processing, or stored?
3. Whether it is sold or shared with research partners, insurers or advertisers.
4. What happens if the company is sold or fails.
5. How to delete it, and whether the answer is written in the privacy policy.

These questions became less abstract in recent years. In 2023 attackers used reused passwords to access the accounts of millions of 23andMe users, and in 2025 the company filed for bankruptcy protection and its assets were later bought by a non-profit research institute. Many people responded by deleting their data at the source. Note that downloading your raw file first keeps your own copy, and deleting the account does not delete that copy.

On our side: the free intolerance test runs entirely in your browser and the file is never sent to us. For paid reports the file is processed to generate the report and then deleted; the details are in our privacy policy. Whichever service you use, keep your own copy in an encrypted folder, do not send it by email, and delete downloads you no longer need.

What to do with it