What i Numbers Instead of rsIDs Mean in Your 23andMe Raw Data

The i numbers in your 23andMe raw data are not damage. They are 23andMe internal marker names, and they are normal. A valid 23andMe file can contain a mix of rsID and i-number lines. If your file has i IDs instead of rsIDs in some rows, judge the file by its header, columns, marker count and no-call rate, not by the i IDs alone.

Why does my 23andMe file have i numbers instead of rsIDs?

23andMe raw data is a plain text file with four tab-separated columns: rsid, chromosome, position, genotype. Most rows use rsIDs such as rs4988235. An rsID is a public identifier from dbSNP, a database of small genetic variants. Some rows have IDs that start with i, like this illustrative row:

i4001234 1 1234567 CT

The i is a 23andMe internal marker name, not a public identifier from dbSNP. The row still has chromosome and position columns, but the build is whatever the file's genome build says; recent 23andMe exports are GRCh37. Some third-party tools match only by rsID and may not use i-number rows. The i number itself is not a public identifier you can look up in dbSNP.

A line starting with i is a real marker, just one that 23andMe names internally instead of by rsID. It does not mean the file is damaged or invalid.

What a valid 23andMe file looks like

A valid download is a .zip containing one .txt. The text starts with comment lines beginning with #. The first line reads # This data file generated by 23andMe. After that, each data line has four tab-separated columns.

A no-call is written as -- and means the chip did not read that position at that line. i-number lines can appear among the rsID lines through the file, not in one broken block. Recent v5 kits list roughly 630,000 markers, and a normal no-call rate is usually well under 5%. The HealthOS DNA file checker reads the file in your browser and reports the company, chip, build, marker count and no-call rate, plus whether the free report can read it.

Open the file in a text editor, not Excel. Excel can re-save with semicolons or merged cells and make a valid file look broken. A text editor shows you the real tabs and the real marker names.

When an i-number line is not the problem

If your file has i numbers, that alone is not a fault. Real problems look different. Check for these instead:

• The file is the PDF or web report, not the raw data download.
• The file is still zipped and a tool wants the .txt inside.
• The download stopped early or ends mid-line.
• The text was opened in Excel and re-saved, changing tabs to semicolons.
• The checker shows a marker count or no-call rate far outside the expected range for your file version.
• The file is unusually short.

None of those are caused by i-number lines. If you see i IDs and the checker shows roughly 630,000 markers for a recent v5 kit and a no-call rate well under 5%, the file is behaving normally.

Will i numbers affect my DNA report results?

It depends on the tool. HealthOS can check whether the free report at https://healthosx.com/dna can read the file. Drop the file at https://healthosx.com/dna/check and the checker reports company, chip, build, marker count and no-call rate, then says whether the free report can read it. Third-party tools that match only by rsID may not use i-number rows. That is a limitation of those tools, not a fault in your file. If a report shows a marker count, compare it with the checker's marker count.

Do you need to convert or fix the i numbers?

No. Do not rename i IDs to fake rsIDs. A fake rsID would not match any public database entry and would only create a new problem. If a third-party tool requires an rsID for every row, it may ignore i-number rows. HealthOS does not ask you to convert or fix i numbers.

A file with i numbers is still a 23andMe raw data file. The format and columns are right, and the genotypes are real.

Drop your file at https://healthosx.com/dna/check to see the company, chip, build, marker count and no-call rate, all in your browser with nothing uploaded. If the file is damaged or truncated, the checker will name the actual problem. If it is just i-number lines, you can stop worrying and use the file. Then see your full report at https://healthosx.com/dna.