HGNC Renamed 27 Human Genes to Stop Excel Date Conversions

ByGeoff Dyers
excel gene: A laboratory workspace with a spreadsheet on a laptop beside a microscope, showing how gene symbols can collide with spreadsheet formatting.

Twenty-seven human genes were renamed in 2019 because Excel kept auto-converting symbols such as MARCH1 and SEPT2 into dates, corrupting gene lists often enough that the HUGO Gene Nomenclature Committee decided the names themselves had to change. HGNC says it updated the affected approved symbols so they would no longer be mistaken for months or dates by spreadsheet software, with examples including SEPT1 → SEPTIN1 and MARCH1 → MARCHF1.

This was not a cosmetic cleanup. HGNC’s own FAQ says the committee changed symbols that “affected data integrity” because spreadsheet auto-formatting errors were a persistent problem in published research and shared datasets, not a one-off annoyance (HGNC FAQ). The weird part is that a naming standard for biology ended up being rewritten around office software behavior. That tells you something about how often messy tools win.

Why the genes were renamed

The reason is simple: some approved human gene symbols looked too much like dates, and spreadsheet programs would silently reinterpret them. HGNC says symbols such as MARCH1 could become “1-Mar,” and symbols in the SEPT family could become September dates when files were opened or pasted into spreadsheet software (HGNC FAQ; HGNC Guidelines).

HGNC formally changed the affected names in 2019 after deciding that leaving them unchanged caused too much downstream damage. In its nomenclature guidelines paper, the committee says it updated symbols that were “known to be problematic,” explicitly citing software auto-conversion, and gives the two flagship examples: SEPT1 became SEPTIN1 and MARCH1 became MARCHF1.

This was meant as a durable fix for future data handling, not just a note telling researchers to be careful. HGNC’s guidelines now say approved symbols should avoid forms that are readily altered by common software, including Excel-style date conversion (HGNC Guidelines).

Which symbols changed

The changed symbols fell into two obvious groups described by HGNC and its guidelines paper: MARCH genes and SEPT genes. HGNC replaced the MARCH root with MARCHF and expanded SEPT to SEPTIN for the affected human genes (Guidelines for Human Gene Nomenclature).

Examples named by HGNC include:

HGNC says 27 approved human gene symbols were changed in total. The committee’s advice to authors is to use the current approved symbols and, where possible, include the stable HGNC ID so old aliases do not create new ambiguity.

If you need to verify a symbol, HGNC provides a Multi-symbol Checker that flags outdated or alias names against the current approved set. That is the practical fix for anyone cleaning old supplementary tables or inherited lab spreadsheets.

What this says about Excel and data handling

A person entering data into a spreadsheet, illustrating how spreadsheet auto-formatting can affect scientific tables.

Excel turns gene symbols into dates because spreadsheets aggressively guess data types. A string like MARCH1 contains a month name plus a number, so software may interpret it as a date rather than literal text; SEPT2 can trigger the same behavior because SEPT is read as September (HGNC FAQ). In other words, the software is trying to be helpful and instead edits the data behind your back.

The broader problem was large, not anecdotal. A widely cited survey in Genome Biology in 2016 found that about one-fifth of papers with Excel gene lists contained gene name errors caused by spreadsheet software. The authors screened supplementary files from genomics papers and found thousands of affected lists across journals and years (Ziemann, Eren, El-Osta, 2016). That is the part that matters: the naming change was not solving a hypothetical edge case.

One useful derived number makes the scale clearer. If roughly 20% of papers with spreadsheet gene lists had these errors, that means about 1 in 5 such papers carried silently mangled symbols into the literature. For a problem caused by default software behavior, that is an absurdly high failure rate.

HGNC’s change helps prevent future corruption, but it does not magically fix historical datasets, old papers, or custom pipelines that still use legacy names (HGNC FAQ). The permanent part is the approved nomenclature going forward; the messy part is everything already published. This is the same class of problem you see whenever tools optimize for convenience first and data integrity second, a pattern that shows up far beyond biology, including the economics of shipping broadly used software fast, as in this piece on Microsoft’s AI cost problem.

For researchers, HGNC’s own recommendation is straightforward: use approved symbols, cite HGNC IDs, and check symbol lists before publication with the committee’s tools (Instructions to Authors; Multi-symbol Checker). That is less elegant than “Excel should have behaved better,” but it is the fix that actually exists.

Key Takeaways

Further Reading

References

Last reviewed: 2026-06