Fraud in scientific publications appears to be on the rise. At least if you can believe all the anecdotes you hear.

By Way of Science You need only look at two excellent websites to get an idea of what's going on out there: Retraction Watch and Abnormal Science.

I enjoy reading both blogs, even though I have great faith in the scientific system. Perhaps even because of that faith. I want to see Science (with a capital S) correct its wayward members and bring the enterprise back onto the "straight and narrow."

If you spend any time looking at either site, you will immediately notice one of the great problems in molecular biology today: image manipulation. By image manipulation I mean copying or altering "bands" (those little smudges of DNA or protein) visualized on gels or membranes. These techniques, including RT-PCR and western blots, are mainstays of most molecular research today in biology, in fields from agriculture to cancer to physiology and beyond.

Compared to the insidious (and also common--I'm looking at you, undergrads I've taught!) phenomenon of plagiarism, image manipulation is both easier and harder to detect. By easier, I mean some of the image enhancement techniques that immediately lay bare all kinds of malfeasance. (Check out some of the comments in this Retraction Watch article, for example.) By harder, I mean simply this: A Google search or use of software such as Turn It In can easily detect plagiarism. Image duplicates usually take a lot longer to detect however.

I was thinking about this this morning in the Responsible Conduct in Research course I am auditing.

What about Image Search?

Unlike common text searches on Google, etc., image search is not yet commonly used. But it has a lot of potential to pick up some kinds of image manipulation (particularly duplicates) in the scientific literature. I hope journals will be proactive and start including this kind of filtering pre-publication for all scientific articles. Might prevent us seeing the cases we're seeing now, where individual scientists or research groups are retracting dozens of papers, each full of duplication, falsification and fabrication.

You can do a simple Google Image Search and it works quite well for normal images. Will it work for gels and blots? I'm sure someone must be working on it. Let's hope so!

You may be asking yourself: What is the government doing about this? It varies a lot by country, but if you want to read about the US government's role in policing science, there's a logical place to start. The ORI website has some great resources, including summaries of prior misconduct inquiries.

On a personal note, I feel strongly about the importance of keeping an eye on science and keeping it honest. Mistakes happen. But I remember as a grad student, a postdoc found a paper in which someone had obviously manipulated an image (to hide "non-specific" bands on a western blot--although to be fair, you had to blow the image way up to be able to see the distortions). Although we thought it should be reportable, he was concerned about retaliation from a big name in the field. I can't blame him, but if I had been in his shoes, I would have reported the incident. For his part, I think he was happy to know at least why he saw two bands on his gel when others had seen only one with the same antibody.

But the literature stayed wrong due to this misconduct. Science needs to stay on top of this and bring its malcontents to justice, or it will continue. Not only that, but "errors" will remain in the literature, poisoning progress in these areas for some period of time.


K.S. MacLea, Ph.D.