They were given names by careful committees. They carried those names in a thousand spreadsheets. And each time a file was opened, the software looked at SEPT1 and saw the first of September.
In 2020, the HUGO Gene Nomenclature Committee renamed 27 human genes to protect them. This memorial preserves what they were called before. Excel was not consulted, out of caution.
Each entry records the fallen symbol and the date Excel believed it to be. The committee reassigned them all to conversion-proof names in the SEPTIN, MARCHF, and DELEC families, where they live quietly to this day.
Read by the spreadsheet as the first through fourteenth of September. All fourteen now carry SEPTIN names. SEPT13, a pseudogene, was given SEPTIN7P2 and does not match its siblings.
Membrane-associated RING-CH ligases, read as the first through eleventh of March. All eleven now carry MARCHF names, the F standing for family, added purely to break the pattern the spreadsheet recognised.
Read as the first of December. The only member of its family, renamed DELEC1. It fell alone and is remembered alone.
Enter a gene symbol, sample ID, or any identifier. The memorial will determine, using Excel's own conversion habits, whether it would survive being pasted into a default spreadsheet.
This is all real. A 2016 study found gene name errors from autoconversion in about one fifth of genomics papers with supplementary Excel gene lists; a 2021 follow-up found the rate had risen to nearly a third. Identifiers like 2310009E13 fared no better, converted to scientific notation on sight.
In 2020 the nomenclature committee formally renamed the vulnerable symbols: the SEPT genes became SEPTIN, the MARCH genes became MARCHF, and DEC1 became DELEC1. It remains one of the only documented cases of humanity renaming parts of its own genome to accommodate a spreadsheet's default settings.
Excel gained an option to disable automatic data conversion in 2023, roughly 36 years after launch. The memorial notes the response time without further comment, as further comment would be undignified in this setting.
Sources visitors may consult: Zeeberg et al. 2004, Ziemann et al. 2016, Abeysooriya et al. 2021, HGNC guidelines 2020. All real papers. The first one is from 2004, which is how long this took.