Photoshopping our way to an amyloid train wreck.

I love Photoshop.  If I do someone's portrait, and there are abnormal skin conditions (for example, too much red or a transient blemish) it is so easy make the person look so much better.  In a landscape photograph, there may be some power lines that can be removed to make the image much more pleasing.  However, I sometimes get grief from people for "cheating" to make the subject look the way they would in different lighting or a few days later.  Or simply for doing manually what your phone, camera, or computer would have done automatically to the image.

But in the world of photojournalism and creating images as part of a scientific or historic record, altering an image, at least manually, is forbidden.  If you were fortunate enough to be the only person who photographed an unexpected event, the wire service isn't going to be willing to buy the image from you without confidence that you didn't alter it.  Fortunately, image editing software often leaves artifacts in the image that give away the fact that it was altered. 

You may have read that the world of Alzheimer's research has been rocked by the revelation of some apparent fraud in the study of the role of amyloids in the disease.  My sister sent me a link to this article in the journal Science which first broke the story, but it's been all over the news since.  The upshot is that, in 2006, a researcher published research purporting to show the role of a previously unknown species (or "oligomer") of the amyloid peptide, Aβ*56 (pronounced “amyloid beta star 56”).  This prompted more than a decade of new research investigating the role of Aβ*56 in Alzheimer's disease.

One problem.  Everyone uncritically took the word of the original researchers regarding their finding -- even though no one was able to show that Aβ*56 actually existed.  I read some comments by some who tried.  When they couldn't find anything resembling Aβ*56, they either just moved on to something else or they were left puzzled.  It didn't seem to occur to anyone that it could be fraud.

The article in Science tells the story of some who doubted the narrative and looked into it.  The researchers had used a technique called the "Western blot" to separate different species of amyloid for individual study.  To apply this technique, researchers create a scan using staining, immunofluorescence, and radioactivity.  This should provide a fingerprint of the peptide or protein being isolated.

In closely examining the Western blot images used to make the claim for the existence of Aβ*56, the investigators found evidence that the images had been manipulated, perhaps with image editing software.  Suddenly the suspicions of a few were finding validation.  The original research (published in the very prestigious journal Nature) had not been seriously questioned up until now.  We will see how the story unfolds from here, but it doesn't look good for the researchers or for the reputation of the University of Minnesota.

If you read Beating the Dementia Monster, you know that I'm something of an amyloid skeptic.  I suspect that we will eventually find out that beta amyloid plays a role in the development of Alzheimer's disease, but it is not a central role.  (But who am I?)  So the first question that jumped into my mind when I read about this was, "Does this kill the amyloid hypothesis?"  Well, probably not.  If Aβ*56 is a species of beta amyloid, then the hypothesis that it plays a special role in Alzheimer's disease is a "species" of the amyloid hypothesis.  If Aβ*56 can go away without beta amyloid persisting, than so can its hypothesis -- probably.  The Science article contains the following observation:

"[Harvard University’s Dennis] Selkoe adds that the broader amyloid hypothesis remains viable. 'I hope that people will not become faint-hearted as a result of what really looks like a very egregious example of malfeasance that’s squarely in the Aβ oligomer field,' he says. But if current phase 3 clinical trials of three drugs targeting amyloid oligomers all fail, he notes, 'the Aβ hypothesis is very much under duress.'”

So ... stay tuned