Incident: A spreadsheet quietly turned names into dates

Module 8 · Incident file · About 30 minutes

What happened

Scientists who study genes share their results as tables: one row per gene, with the gene's short name in the first column. Many of those tables pass through a spreadsheet on the way.

In 2016 Mark Ziemann, Yotam Eren and Assam El-Osta published a study of this in the journal Genome Biology, "Gene name errors are widespread in the scientific literature". They found that opening such data in a spreadsheet had changed some gene names into dates. A gene called SEPT2 is a name, not a day in September, but the spreadsheet read it as one. Of the papers they checked, about a fifth had the error.

Nobody had typed a wrong value. The names were right when the file was opened, and wrong when it was saved. The damage sits in published work, where other people may reuse it. In the end the problem was also handled at its source: the gene names were later renamed.

This file keeps to those facts.