Thursday, January 04, 2007

Errors in scientific papers, caused by software

Science (22 December 2006) reports the scientific breakthroughs of the year, which makes interesting reading, but also has this news item, reporting the retraction of several crystallography papers, which contained errors caused by the data analysis software flipping two columns of data. This resulted in structures being inverted. One paper, published in 2001, has been cited over 300 times.

This caught my eye as I have a long interest in scientific publishing, and in retractions and related areas. But it is also interesting because the things that were being analysed were proteins from E. coli, Vibrio cholera and S. typhimurium.

The full news piece is at: http://www.sciencemag.org/cgi/content/summary/314/5807/1856 and has more detail about how the error was discovered.

No comments: