Do I Need to Unzip My 23andMe or AncestryDNA Raw Data Before Uploading It to Another Site
A 23andMe or AncestryDNA download is a .zip holding one .txt. Some sites take the zip, some want the .txt, and the real damage comes from re-zipping or saving it in a spreadsheet.
No, not by default. A 23andMe or AncestryDNA raw data download arrives as a .zip holding a single .txt, and many DNA sites accept that zip exactly as it came. Unzip only when a tool says it wants the .txt, extract it once, and leave the text untouched.
Why the same zip works on one site and fails on another
It depends on the site, and the site's upload box usually says which. A zip is a wrapper: the text inside has been compressed, and something has to decompress it before the rows can be read. A tool that handles zips does that for you. A tool that does not sees a block of compressed bytes (a zip begins with the letters PK) and finds no rows, so it calls the file empty or unreadable.
So the same file can work on one site and bounce off another. Nothing is wrong with your DNA or your download. One program opens the wrapper and the other does not.
| Situation | What to upload |
|---|---|
| Site says .zip is accepted | The zip as downloaded |
| Site says it wants the .txt | The one .txt, extracted once. A zip here may read as empty or unreadable |
| Site does not say | Try the zip, then the .txt |
| You opened and saved it in a spreadsheet | Neither. Go back to the original zip |
Go by what the upload box says, not by habit.
How to unzip it properly: on Windows, right-click the zip and choose Extract All. On a Mac, double-click the zip (or Safari already did it for you). On an iPhone, tap the zip in Files.
Mac and iPhone: the zip may already be gone
On a Mac, Safari's "Open safe files after downloading" setting is on by default in recent versions. By default, Safari unzips a downloaded .zip and moves the zip to the Trash, so look for the .txt in Downloads first, and check the Trash if you want the original zip back. Chrome and Firefox on a Mac leave the zip as it is, so there is nothing to look for in the Trash if you use one of those. Having only the .txt is fine, as long as it is untouched. On an iPhone, tapping the zip in the Files app extracts it in place, so you may already have the .txt there too.
What is actually inside the zip
The 23andMe .txt starts with comment lines beginning with #, the first reading "# This data file generated by 23andMe". After that come four tab-separated columns: rsid, chromosome, position and genotype. A recent v5 kit lists roughly 630,000 markers on genome build 37. A genotype of -- is a no-call, a position the chip did not read, and a small share of those is normal.
AncestryDNA's download is also a zip with one text file inside. Its layout differs: the two letters sit in separate columns, and a no-call is written as 0. The file is not a 23andMe file in a different box, so a tool that expects the 23andMe layout can refuse it even after you unzip it. That is a format problem, not a zip problem, and the free AncestryDNA to 23andMe converter fixes the layout. It runs in your browser, so the file is never uploaded.
How the file gets damaged before upload
The damage usually happens before the file reaches any site. The table below is the one place that lists all four ways it goes wrong, what each does, and what it tends to look like when a tool complains. The spreadsheet case gets its own section next, because people rarely notice it.
| Habit | What it does to the file | What you may see |
|---|---|---|
| Interrupted download or extraction | The text stops partway, sometimes mid-line | A marker count far below expected (about 630,000 for a 23andMe v5 file) or a complaint about a cut-off line |
| Re-zipping a folder | The .txt ends up one level down, and a Mac may add a hidden __MACOSX folder with stray files | No genotype rows found, or the wrong file read |
| Renaming .zip to .txt, or adding .txt to an extracted file | Still a zip with a misleading name, or named name.txt.txt when extensions are hidden | Empty, unreadable or unrecognised file |
| Saving in a spreadsheet as .csv or .xlsx | Tabs become commas, or it is no longer text (see the next section) | Wrong format, or no rows found |
Scroll the table sideways for the rest →
Based on how the formats work, not on any one site. Wording of real error messages varies by tool.
The spreadsheet mistake
This is the one people do not notice. Opening the .txt in Excel or Numbers is harmless. Saving it is not. Save as .csv and the tabs become commas. Save as .xlsx and it is no longer a text file at all.
A chip file has about 630,000 rows, under Excel's limit of 1,048,576, so nothing visibly breaks. The sheet looks perfect. But a tool that splits each line on tabs can reject it.
A zip is a wrapper, not a different file: hand a site whichever one it asks for, and never one you have edited.
Can I open it to look without ruining it?
Yes, in a plain text editor (Notepad, or TextEdit in plain text mode), and close it without saving. You should see the # lines first, then one marker per line. If you see a wall of strange characters, you opened the zip, not the text.
A sensible habit: keep the original zip somewhere untouched as your master copy. Extract a working copy from it whenever a site wants the .txt, and if anything goes wrong, extract again rather than repairing a damaged one.
When the zip is not the problem
Site rules differ and change. A correctly handled file can still be refused for reasons that have nothing to do with zipping. The tool may want a different company's layout, a different genome build (consumer chips report on GRCh37, many sequencing providers on GRCh38), or a cap on file size.
If the error message does not say which of these it is, guess less and look at the file itself.
The free DNA file checker reads a .zip as it is, so you can drop the file straight from your downloads. It tells you whether it is readable, which company and genome build it is, how many markers and no-calls it holds, and why another tool may have refused it. It runs in your browser, so the file is never uploaded.
For the full free report, go to healthosx.com/dna, which reads the zip as downloaded.
New to raw DNA files? Start with the full guide: 23andMe Raw Data Interpretation Free in Your Browser
Written by
HealthOSX Research