Working in this field, I think a lot of the issue is publishers often require small datasets in xls format for paper submission. So someone in the chain (corresponding author, editor, journal staff member etc.) opens the file in excel and saves it again.

Bioinformaticians are not doing their analyses in excel.