qpAdm from your 23andMe, AncestryDNA or MyHeritage raw data: what works, what to expect

Raw data files from 23andMe, AncestryDNA and MyHeritage are among the most common inputs we see for Ancient Ancestry, our qpAdm analysis, and all three are good ones. None of them is a special case: qpAdm does not care which company genotyped you, only which positions your file carries and how many of them survive the merge with the ancient reference panel. The company still matters in three practical ways: how you get the file out, how it is laid out, and which chip it was read on. The chip is the one that sets your coverage. This guide takes each company's file in turn, then follows it through an order to the report, and it tries to be exact about what the free file check can tell you, what coverage changes in the result and what it never changes. If you want the prices first, they are in what a qpAdm analysis costs.

What each vendor's file looks like#

All three exports are the same kind of object: a table of several hundred thousand positions where people commonly differ, with your two alleles at each. Keep the file exactly as the vendor delivers it. Our uploader reads .zip, .gz, .txt and .csv files up to 50 MB and recognises the format by its content rather than its name, so there is nothing to unpack, rename or convert. FamilyTreeDNA and LivingDNA exports are accepted the same way, and how to download your raw DNA data has every vendor's steps, checked against their own help pages.

23andMe#

23andMe does not release the file instantly. In your account, open Settings, scroll to 23andMe Data, and choose Download Raw Data. The site asks you to confirm the request and then sends an email when the file is ready (23andMe says this usually takes about a week and can take up to thirty days); follow the link in that email to fetch a .zip holding a single .txt file.

Every line of that file is one position: an rsID, a chromosome, a base-pair position and your two alleles, written in the forward orientation of the GRCh37 build. Positions the chip could not call are written as --. These are not errors; they are simply absent from any analysis, which is why a file's usable marker count is always lower than its line count.

23andMe has shipped several chip generations, and the generation decides how many positions your file shares with the ancient panel. Roughly:

• v3 (2010 to 2013) sat on a large Illumina OmniExpress-derived design of approximately 960k positions, the most of the three, and it usually overlaps the ancient panel best.
• v4 (2013 to 2017) moved to a custom design of around 600k positions and a different selection philosophy, so a v4 file often overlaps the ancient panel less than a v3 file does despite being newer.
• v5 (2017 onward) is built on Illumina's Global Screening Array (GSA), approximately 640k positions. Its overlap with the ancient panel is respectable but not the largest of the three.

The version is usually part of the file's name: a v5 download is named genome_Your_Name_v5_Full_ followed by a date stamp. These numbers are approximate and the designs have been revised within versions, so treat them as orientation rather than a promise.

AncestryDNA#

Ancestry, too, sends the file by email rather than from the results page. Open Settings from your DNA test page, choose Download DNA data, and confirm with your password or a code sent by email. Ancestry then emails you a link (its help page says the email can take up to 24 hours), and that link is what actually delivers the file, a .zip containing a single .txt. The link works once and expires after a week, and only the latest request's link is valid, so if you leave it too long, request the file again.

AncestryDNA has used two chip designs:

• v1 (2012 to 2016) was an Illumina OmniExpress-derived array of approximately 700k positions.
• v2 (from 2016) moved to a custom design of roughly 650k to 700k positions that kept only part of v1's positions, so a v1 and a v2 file from the same person meet the ancient panel at different sites.

The version is written in the file's header comments, in a line reading #Data was collected using AncestryDNA array version: V2.0 (or V1.0). Ancestry has revised the design within v2 more than once, so here too the figures are approximate.

Two more things make an AncestryDNA file look different from a 23andMe one. The first is the columns. Ancestry writes each position as rsid, chromosome, position, allele1 and allele2, tab-separated, with the two alleles in separate columns rather than joined, and an uncalled position as 0 0. Some tools that expect the joined 23andMe form choke on this; ours joins the columns itself.

The second is the names. An rsID is only a label, and the name a chip file gives a site is not always the one the ancient panel uses for it. The merge does not match on names: it pairs your positions with the panel's by chromosome and GRCh37 position and reconciles the alleles, so a name that differs from the panel's is not a reason for a position to drop out. Two files with the same line count can still end up with different coverage, because what counts is which positions they carry, not how many lines they have.

Chromosomes 23, 24, 25 and 26 in an Ancestry file mean X, Y, the pseudoautosomal region and mitochondrial DNA; qpAdm uses the autosomes (1 to 22) only. None of this needs any action from you. It is described here so that a strange-looking file, or a warning from some other tool, does not make you think the file is broken.

MyHeritage#

In your MyHeritage account, open the DNA menu and choose Manage DNA kits. Each kit has a three-dots menu on the right; open it, choose Download, and agree to the terms. MyHeritage then emails you a link, valid for 24 hours, and once you enter your password on the page it opens, you get a .zip holding a .csv. Only the kit's manager can download it, and the download works from a computer or an Android device but not from an iPhone or iPad.

The CSV starts with a block of comment lines describing the kit, then a header row, then one quoted, comma-separated line per position: the rsID, the chromosome, the GRCh37 base-pair position and your two alleles joined together, AG for example. Uncalled positions are written as --. The commas alone are enough to make some tools reject the file; ours reads it as it comes.

Which chip read your kit depends on when it was processed. Kits read before 2019 ran on an Illumina OmniExpress-based chip of approximately 700k positions; in 2019 MyHeritage moved to a custom chip based on Illumina's Global Screening Array, the same family as 23andMe's v5. The two carry different positions, so they meet the ancient panel at different sites.

The low-pass sequencing export. In October 2025 MyHeritage announced that its kits would move from the chip to low-pass whole-genome sequencing (guaranteed for kits reaching its lab from January 2026; kits processed earlier are not re-run). The export comes in the same CSV shape as the chip file, with one difference that matters: it carries autosomes and the X chromosome only, with no Y-chromosome rows and no mitochondrial rows.

For qpAdm this is harmless, because qpAdm uses the autosomes only, and the export is read like any other MyHeritage file. It does matter for the optional paternal (Y-DNA) and maternal (mtDNA) haplogroup add-ons, 10 EUR each, which have nothing to read in it. The file check shows zero Y and mitochondrial rows for such a file, and the checkout will not sell either add-on for a file that lacks the rows, so you cannot pay for a haplogroup the file cannot give. If you want one, it has to come from a file that carries those rows, such as a chip export from before the switch.

Will my file work? The free file check#

Before ordering, run the file through the free Raw DNA File Check. It needs a free account with a verified email; it reads the file on our servers and deletes it as soon as the check finishes, so nothing is ordered and nothing is kept. It reports the format it detected, how many rows the file holds on each chromosome (autosomes, X, Y, the pseudoautosomal region and mitochondrial DNA), and one verdict per analysis, qpAdm included, with the reason in plain language.

That verdict is the honest answer to "will my file work": a readable export with a normal autosomal count can run, and a truncated or edited file is caught before you pay. What the check does not measure is how much of your file overlaps the ancient panel, because that needs the reference data only the merge loads. The only number that matters for your model is the one the merge produces, and it is printed in your report as the merged SNP count, one of the model's fit figures.

What an Ancient Ancestry order does with the file#

Once you order, your genotypes are converted, filtered of indels and strand-ambiguous positions, and intersected against the Allen Ancient DNA Resource (AADR) v66 with Poseidon's trident. The result is a merged sample in which your file and more than 23,000 ancient and modern reference samples are read at exactly the same positions. From that point the vendor name is gone; your sample is just a target.

Candidate sources come from a pool curated for your declared origins and for each era before any model runs. Our analyst tests models from that pool (automated searches included) and re-tests the survivors, and a person reviews the candidates, chooses and approves the published model. The pool comes first for a reason: how far a fitting model can be trusted depends above all on the search's pre-study odds, the ratio of plausible to implausible models among those tested (why that matters).

Each model is a small set of ancient source populations and a set of distant outgroups, and each qpAdm run returns a weight, a standard error and a Z-score per source and a single p-value for the whole model. Your report carries one model per era, for two eras (Hunter-Gatherer and Neolithic Farmer; Classical Antiquity), and its Your model view shows each model's record: seven fit figures (p-value, chi-square, degrees of freedom, f4 rank, minimum SNPs per f4, merged SNPs and jackknife blocks) under a verdict against our bar; a Sources table giving each source's panel label, sample count, share, weight, SE, Z and 95% CI; and the right (outgroup) populations with their sample counts. The nested models and the rank test are what the analyst checked before publishing; they are not printed, and there is no downloadable record file. Maps and two rendered videos, a Globe film and a Populations film, each covering every era, come with it.

How those numbers are read is set out in How to read qpAdm results, and every figure of the record in Every number in your qpAdm report.

What changes with coverage: the standard errors#

The standard error on each weight is bounded by how many positions the model could use, and your file's merged count sets that bound. More surviving positions give tighter errors, fewer give wider ones, and no analyst effort can pull an error below what the file allows. This is where the chip shows: a 23andMe v3 file will usually produce tighter errors than a v4 file from the same person, and AncestryDNA's v1 and v2, or MyHeritage's chips before and after 2019, meet the panel at different sites.

Our publish bar is the same for every file and every tier: p > 0.05, and for every source in every era, |Z| > 3 and a standard error below 0.10. It is a house convention, not a validated criterion, and what it guarantees is narrow: with every weight positive, |Z| > 3 places each weight more than three standard errors from 0 and from 1. When a file's coverage leaves the errors above that line, the report does not hide it: each era's model is marked as meeting the bar or sitting below it.

What does not change: the p-value logic#

Coverage does not change what a p-value means. A passing model is one the data could not reject: it describes your ancestry as a composition relative to the sampled groups, not proof that those groups mixed. A failing one is rejected. That logic is identical whether the target came from a 23andMe v3 chip, an AncestryDNA v2 chip, a MyHeritage low-pass export or a whole-genome VCF. Lower coverage makes the test less able to distinguish between close alternatives, which shows up as wider errors, not as a different kind of answer. If you have ever seen a percentage breakdown that never fails, that is a different method entirely; the difference is laid out in qpAdm vs Global25.

Ordering#

The four tiers run from 29.99 to 59.99 EUR, paid once, and every tier gets the same report, maps, videos, model record and publish bar (what each tier costs). Upload the same file you ran through the check on the buying page, tell us which regions your family comes from if you know (that step is optional), and the merge starts on our infrastructure. A qpAdm report is reviewed work, so it does not render the moment you pay.

If you have had your whole genome sequenced, a single-sample .vcf or .vcf.gz on GRCh37 or GRCh38, up to 1 GB, is accepted instead through the 10 EUR whole-genome upload, and a sequencing file usually covers more of the panel than any chip.

Afterwards: the Ancestry Workbench#

Once your report is published, you can open its model in the Ancestry Workbench straight from the Your model view, and your merged sample attaches to the Workbench for free, so nothing has to be prepared again. There you choose sources and outgroups and run your own qpAdm models on your own sample, paying per completed run from a prepaid balance (from 9.99 EUR) that never expires. Most models will be rejected, which is the method working; a completed run is charged whether its model passes or not, and a run that fails to complete is never charged. You can also download the merged EIGENSTRAT bundle the report was computed from and re-run the published model in ADMIXTOOLS 2 on your own machine. The walkthrough is in Run your own qpAdm models.

And if you want to feel a rejection before ordering anything, the AdmixTools 2 Lab runs real qpAdm over the AADR reference panel on our servers, free with an account. It cannot use your sample, but it teaches what a failing model looks like, which is the most useful thing to know before reading your own.

A short summary#

1. Download the file: 23andMe from Settings, 23andMe Data; AncestryDNA from Settings, Download DNA data; MyHeritage from DNA, Manage DNA kits. Wait for the email and keep the zip as it comes.
2. Run it through the free file check and read its qpAdm verdict.
3. If it is a MyHeritage low-pass export, remember it has no Y or mitochondrial rows, so it cannot carry a haplogroup add-on.
4. Order from the buying page; every tier returns the same report.
5. Read the report with the p-value, errors, merged SNP count and outgroup list in view, not the percentages alone, then open the Ancestry Workbench if you want to run the alternatives yourself.

Terms used here are defined in the glossary.