Excel’s date auto-formatting has corrupted hundreds of genetics studies for 15 years

In the summer of 2020 the Human Genome Organisation Gene Nomenclature Committee (HGNC) disclosed a plan to rename 27 human genes whose official symbols are routinely mistaken for dates by Microsoft Excel’s auto-format feature. The change aims to stop a pervasive error that has distorted the results of about one-fifth of genetics studies worldwide. A 2016 analysis of over 3,500 papers identified 704 instances where Excel had altered gene names such as MARCH1 into a calendar date, compromising the underlying data.

Although the flaw was first reported in 2004, Microsoft did not provide a way to turn off the automatic conversion until a 2023 update. HGNC’s Elspeth Bruford speculated that the limited impact on Excel’s overall user base delayed a fix. The rename replaces names like MARCH1 with MARCHF1 and SEPT1 with SEPTIN1, eliminating the ambiguity for future research.