The short answer: FamilyTreeDNA does not and cannot give you Global25 coordinates. No testing company does — not even one as focused on deep ancestry as FTDNA. What FTDNA gives you is a raw autosomal export from Family Finder, and that file is what the independent Eurogenes Global25 service turns into a coordinate row. This guide covers the FTDNA-specific half of the route; the full background is in the main guide.
Step one: download your Family Finder raw data#
In your FTDNA account, open Family Finder → Download Raw Data and choose the autosomal build
export. That is the file Global25 needs: autosomal genotype calls, served as a compressed .csv.
The download page offers more than one file, and the choice matters. FTDNA publishes the same genotypes against two reference builds, Build 37 and Build 38, each in a concatenated form that holds every chromosome in one file. Prefer the Build 37 concatenated export: most third-party ancestry tools grew up on build 37 positions, and a build 38 file can be misread by software that assumes the older coordinates. The Build 38 export exists for people who need it for other pipelines; for this route it is the wrong door. The concatenated file arrives as a compressed .csv.gz, and the portal takes it as delivered, so there is no need to decompress it.
If you cannot find the page, the landmarks are: sign in, open the Family Finder section of your dashboard, and look for a Download Raw Data link in the results area or in the kit's settings menu. FTDNA may ask you to confirm your password before the files are released. If the option is missing altogether, the usual reason is that the kit is a Y-DNA or mtDNA only order, or that Family Finder results have not yet completed.
One point of frequent confusion: FTDNA sells several distinct products, and only Family Finder produces the autosomal file this route needs. A Y-DNA or mtDNA product (Y-37, Big Y, mtFull) tests a different part of the genome and its exports cannot become a Global25 coordinate — Global25 is computed from autosomal data. Those products answer lineage questions instead, which is a different (and complementary) thing.
Check what the file holds#
Marker coverage differs by chip generation even within one vendor, and coverage decides which analyses a file can support. Before sending the export anywhere, our free raw DNA file check parses it, reports the detected format and counts usable markers per chromosome — nothing is ordered and nothing is stored.
File format notes#
The Family Finder export is a comma-separated text file with a header row and four columns: RSID, CHROMOSOME, POSITION and RESULT, one marker per line, with the genotype given as two letters. It is the same information a 23andMe or AncestryDNA file carries in a different column layout, and every tool on the Global25 route reads it without conversion. FTDNA has used more than one genotyping array over the years: Family Finder kits processed since around 2019 run on an Illumina Global Screening Array based chip, while older kits were genotyped on an earlier array with a different marker selection. Both produce valid raw files, but their overlap with the set the Global25 service projects on differs, which shows up as slightly different fit quality rather than as different ancestry.
Step two: the file becomes a coordinate#
Global25 coordinates are issued by Davidski's independent Eurogenes Global25 service — the G25 Requests portal at g25requests.app — not by FTDNA, and not by us. Upload the raw file there, pay their €15 per-kit fee, and receive your row; their site states 2–7 days. The process and terms are the service's own and have changed more than once, so their posting is the only authority on the details. You will receive scaled and unscaled forms of your row: keep both, keep them labelled, and never mix the two in one comparison.
Submitting the file to the portal#
Upload the concatenated file as FTDNA delivered it, pay the per-kit fee and wait for the email with your row. Give the kit a label you will recognise, because it becomes the first field of the coordinate row and travels with it into every tool you paste it into. When the row arrives, run it once through the authenticity check before building on it; that confirms the copy you pasted is the copy you were sent.
If the file is rejected#
Rejections at upload are nearly always packaging. Check that you uploaded the concatenated file rather than a per-chromosome export, that it is the Family Finder autosomal file rather than a Y-DNA or mtDNA result, and that it is the original .csv.gz rather than a copy re-saved from a spreadsheet, which can strip the header or change the delimiter. A file much smaller than a normal export is an interrupted download; fetch it again. If you uploaded the Build 38 file, try the Build 37 one. Past that, the portal's own posting is the authority on what it currently accepts.
Common problems with FTDNA rows#
The usual FTDNA-specific issue is marker overlap. A GSA-era Family Finder file shares fewer positions with the projection set than a 23andMe v5 file does, so expect fit distances in the admixture calculator to sit slightly higher than a 23andMe row for the same person would produce, and the last few dimensions to carry a little more noise. The row is still official and every comparison remains valid; the effect is a property of the input, not an error, and it does not change which populations rank closest. The second issue is the build mix-up described above: if a tool refuses the file or reports far fewer usable markers than expected, check that you exported Build 37. And if your row's nearest population on Earth sits past roughly 0.08, suspect the paste or the scaling before you suspect your ancestry.
Rather not handle the request yourself? Upload this same file with an Ancestrify Global25 analysis and, with your consent, we obtain your official coordinates for you from that very service (+€15, typically a few days) — your full analysis runs the moment they arrive, and the row stays yours to keep.
Avoid the simulated shortcut: converters that fabricate a G25-shaped row from another calculator's output describe the conversion, not your genome. Our authenticity check reads a row's numeric fingerprint whenever provenance is in doubt.
Step three: use the row#
With a real coordinate row, all of this is free, in your browser, no account:
- Rank your closest ancient and modern populations, era by era, in the G25 distance calculator.
- Model your ancestry against curated source panels in the Global25 admixture calculator.
- Plot yourself on ancient-DNA PCA views in the G25 PCA viewer.
For the full worked report — distances, admixture and PCA across six eras with cinematic exports — see our Global25 analysis. And if you would rather submit the raw file itself and get a formal, testable ancestry model with p-values and standard errors, that is our qpAdm analysis, which takes the raw file directly.
Frequently asked questions#
Does FamilyTreeDNA provide Global25 coordinates?#
No. FTDNA reports its own myOrigins breakdown and, for the relevant products, Y-DNA and mtDNA results. None of that is a Global25 row. Coordinates come only from the independent Eurogenes service, from the Family Finder raw file.
Which FTDNA download should I use, Build 37 or Build 38?#
Build 37, concatenated. Most third-party ancestry tools expect build 37 positions, and the concatenated file holds every chromosome in one place. Keep the Build 38 file only if another pipeline specifically asks for it.
Can I use my FTDNA file in Ancestrify's free haplogroup finders?#
Not at present. The free Y-DNA clade finder and mtDNA finder read 23andMe, AncestryDNA and MyHeritage exports (the clade finder also takes a VCF); Family Finder files are not among the supported formats, and a Family Finder file is autosomal data in any case. FTDNA's own Y-DNA and mtDNA products answer those questions directly.
Terms used here are defined in the glossary.
