Monday, January 30, 2017

Microsoft Excel linked to high number of errors in genetic scientific literature

Studies have found that Microsoft Excel incorrectly converts gene names to dates. Approximately one fifth of leading journals on genomics with supplementary Excel gene lists contain erroneous name conversions compromising the integrity. For example, SEPT2 (Septin 2) would be automatically "corrected" to September 2. The problem is there seems to be no way to disable the auto-date formatting in Excel, so researchers have to manually enter the text, but this is time consuming and makes human error more probable as most won't always remember to do this or may not even be aware of Excel's auto-formatting.  Australian researchers analyzed thousands of published papers on genetics in leading scientific journals such as Nature, Science,  and PLoS One, and found that roughly 20% of these papers have errors in their supplementary files containing the genes used.
 The scary part is that this was found in a paper published over a decade ago, and the problem still persists.

Apparently genetics is not the only field subjected to such errors. Harvard economists found Excel spreadsheet errors for calculating GDP, and over-inflating projected numbers. Excel isn't the only culprit, LibreOffice Calc and Apache OpenOffice were found guilty as well of having the same auto-formats.  However, there was only one program found to contain none of the errors and that is none other than Google Sheets. Even Science Mag acknowledges one in five papers contain errors. 

This is really troubling as it seems to me like a virus has penetrated the scientific world where technology induced errors are becoming rampant and since the modern man's life is essentially dominated by such technology, it gives rise to concern on how to not only correct this problem, but to prevent them in future.

No comments:

Post a Comment