AncestryDNA has the largest customer base of any consumer test, so its raw file is the one we are asked about most. The short version: it works well as a qpAdm input, its coverage against the ancient panel is among the better ones for a chip file, and it has one formatting quirk that our uploader handles for you but that you should know about if you ever move the file between tools. This guide walks the AncestryDNA route from the download menu to the report and to the Model Lab afterwards. If you already have the file, the tiers are on the buying page.
Step one: download the raw file#
Ancestry does not expose the file on the results page. Open Settings from your DNA test page,
find Download DNA data, and confirm with your password. Ancestry then sends an email; the
download link inside it is what actually hands you the file, a .zip containing a single .txt.
The link expires after a while, so if you leave it a few days you may have to request again.
Keep the zip as it comes. Our uploader accepts .zip, .gz and the unpacked .txt, and it
identifies the file by its content rather than its name, so nothing needs renaming.
v1 and v2: which chip you have#
AncestryDNA has used two chip designs:
- v1 (2012 to 2016) was an Illumina OmniExpress-derived array, approximately 700k positions, with solid overlap with the 1240k capture set used for most ancient genomes.
- v2 (2016 onward) moved to a custom design of approximately 680k positions. A large share of the v1 positions were kept, and the overlap with the ancient panel is good, though not every position carried is one the ancient samples were sequenced at.
These figures are approximate, and Ancestry has revised the design within v2 more than once. The version is written in the header comments of the file, and the free file check reads it for you. What decides your model is not the version but the count that survives the merge.
The Ancestry quirk: allele encoding and rsIDs#
Two things make an AncestryDNA file slightly different from a 23andMe or FamilyTreeDNA export.
First, the columns. Ancestry writes each position as rsid, chromosome, position, allele1,
allele2, tab-separated, with the two alleles in separate columns rather than joined. Uncalled
positions carry 0 0. Some tools that expect the joined 23andMe form choke on this; ours reads
both layouts.
Second, the identifiers. Ancestry files carry a noticeable number of positions with rsIDs that are either internal to the chip design or absent from the public dbSNP releases the ancient panel is keyed on. Those positions cannot be matched by name. We match on chromosome and GRCh37 position rather than on rsID where the name fails, so the loss is small, but it is one reason two files with the same line count can end up with different coverage after the merge. Chromosomes 23, 24, 25 and 26 in an Ancestry file mean X, Y, the pseudoautosomal region and mitochondrial DNA; qpAdm uses the autosomes 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.
What to check first: the free file check#
Before ordering, run the zip through the free Raw DNA File Check. It detects the format and chip version, counts usable markers per chromosome, and returns a coverage verdict for qpAdm: how many of your positions intersect the Allen Ancient DNA Resource (AADR) v66 panel that every paid model is run on. Nothing is ordered and nothing is stored.
That verdict is the number to read. A valid file with a small overlap will produce standard errors that cannot clear the publish bar, and it is far better to learn that for free than after paying.
What a qpAdm order does with the file#
After you order, your genotypes are converted, filtered of indels and strand-ambiguous positions, and intersected against AADR v66 with Poseidon's trident. What comes out is a merged sample in which your file and roughly 23,265 reference samples are read at exactly the same positions. The vendor is now irrelevant; your sample is a target and nothing else.
An analyst then composes models by hand: a small set of ancient source populations, a set of distant outgroups, and a qpAdm run that returns a weight, a standard error and a Z-score per source plus one p-value for the whole model. The report publishes all of it for each of two eras, with the complete outgroup list and the full run record behind it. Reading those numbers is covered in How to read qpAdm results.
What changes with coverage: the standard errors#
The standard error on every weight is bounded by how many positions the model could use, and an Ancestry file's coverage is what fixes that bound. More surviving positions means tighter errors; fewer means wider, and nothing an analyst does can pull an error below what the file allows. The 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. A file that cannot reach it does not get a published model, and the file check is where you find out in advance.
What does not change: the p-value logic#
Coverage does not alter what a p-value means. A passing model is one the data could not refute; a failing one is refuted. That is true for a v1 file, a v2 file and a whole-genome VCF alike. Lower coverage makes the test less able to tell close alternatives apart, which appears as wider errors, not as a different kind of answer. A breakdown that never fails is a different method; the distinction is set out in qpAdm vs Global25.
Ordering#
The buying page lists the four tiers, from 29.99 EUR. The publish bar and the report are identical at every tier; what a deeper tier buys is a longer search past the first model that clears the bar, so that more alternatives have been tried and rejected before one is shown to you. Upload the same zip you ran through the file check, pick the region you want the model scoped to, and the merge starts on our infrastructure. The report is reviewed work and does not render the moment you pay.
If you have since sequenced your whole genome elsewhere, a .vcf or .vcf.gz up to 1 GB is
accepted instead through the 10 EUR whole-genome upload; a sequencing file usually covers more of
the panel than any chip does.
Afterwards: the Model Lab#
Once the report is published, a one-time 10 EUR unlock opens the Model Lab. Your Ancestry genotypes are already merged into the panel, so you choose sources and outgroups and run your own models, up to 100 per rolling 24 hours. Most will fail; that is the method doing its job. You can also download the exact EIGENSTRAT bundle the report was computed from and reproduce everything in ADMIXTOOLS 2 on your own machine. The walkthrough is in Run your own qpAdm models.
To feel a rejection before ordering anything, the free AdmixTools 2 Lab runs real qpAdm over the public reference panel in your browser. It cannot use your sample, but it shows what a failing model looks like, which is worth knowing before you read your own.
A short summary#
- Settings, Download DNA data, confirm, wait for the email, keep the zip.
- Run the zip through the free file check and read the coverage verdict.
- Order the qpAdm analysis at whichever tier matches how hard you want the search to be.
- Read the report with the p-value, errors and outgroup list in view, not the weights alone.
- Unlock the Model Lab if you want to run the alternatives yourself.
Terms used here are defined in the glossary.
