Humanity renamed genes because of Excel - Davide Mornatta
← Humanity renamed genes because of Excel
In 2004, bioinformaticians started noticing that, when processing tabular datasets, some gene names<br>were being changed inadvertently into non-gene names.
Upon investigation, the culprit turned out to be<br>Excel's automatic type conversion.<br>When Microsoft Excel imports or pastes data, it automatically guesses the data type of each cell. A string like<br>"SEPT2" gets interpreted as 2-Sept.<br>This is convenient for most users, but for biologists it's a complete disaster.
Imagine a molecular biology lab processing gene expression data.<br>One column has:
TP53
BRCA1
SEPT2
MARCH1
GATA3
It gets transformed into:
TP53
BRCA1
2-Sep
1-Mar
GATA3
When the spreadsheet is later exported, downstream software no longer recognizes those genes.<br>Instead of analyzing SEPT2 or MARCH1, the software simply sees dates and may ignore them, silently discarding<br>important data.<br>Researchers in 2016 scanned the leading genomics journals and revealed that<br>approximately<br>one-fifth of papers with supplementary Excel gene lists contained erroneous gene name conversions.
This also meant that a large number of widely used datasets contained corrupted data. Some of them probably are<br>still used without awareness of the problem or without being fixed.
The funny turn of the story is that an uncommon decision was taken. To address the problem, in 2020 the HUGO<br>Gene<br>Nomenclature Committee decided to<br>rename dozens of<br>official human gene symbols<br>so they could no longer be mistaken for dates.
Science adapted to the program.
Symbols like SEPT1 became SEPTIN1 to avoid the problem.<br>And there is more, because, due to the above mentioned fact that millions of existing datasets still contained<br>the old errors, researchers have created dedicated tools to repair them.<br>Two examples are<br>"Escape Excel"<br>(2017), which prevents Excel from auto-converting problematic strings before import, and<br>"Gene Updater"<br>(2022), which automatically converts corrupted dates like "2-Sep" back into the correct modern gene<br>symbols.
Jumping to the current times, the fix mentioned on<br>HGNC FAQ<br>page<br>remains changing the cell formatting to text before importing the data.
TLDR: Humanity permanently changed some gene names because they did not parse well on Excel.
References
Mistaken<br>Identifiers: Gene name errors can be introduced inadvertently when using Excel in bioinformatics<br>(Zeeberg et al., 2004)
Gene name errors<br>are widespread in the scientific literature<br>(Ziemann, Eren & El-Osta, 2016)
Escape Excel<br>(2017)
Guidelines for<br>Human<br>Gene Nomenclature<br>(Bruford et al., 2020)
Gene Updater<br>(2022)
HGNC FAQ —<br>Excel is corrupting some of my gene symbols
Gene names auto-converted to dates in Microsoft Excel<br>(Microsoft Q&A)