Sometimes, retractions happen months, or even years, after another researcher spots problems in a paper. But when it’s a Nobel Prize winner who finds the error, things might move more quickly.
In the case of a recent paper in the Proceedings of the National Academy of Sciences, the retraction happened within six weeks. Here’s the notice for “Voltage sensor ring in a native structure of a membrane-embedded potassium channel,” by Liang Shi, Hongjin Zheng, Hui Zheng, Brian A. Borkowski, Dan Shi, Tamir Gonen, and Qiu-Xing Jiang, which first appeared online on February 11:
The authors wish to note the following: “The contrast of our final projection map was inverted, so that we interpreted the background density rather than the actual protein density in terms of structural features of the potassium channel-Fv complex. In addition, we indexed the 2D crystals with unit cell parameters of a = b = 175 Å, while the correct indexing would be a = b = 124 Å. Given these analysis errors, the resulting density map and our interpretation of the structural features are not correct. Accordingly, we would like to retract this paper. We acknowledge Yoshinori Fujiyoshi, Rod MacKinnon, Kazutoshi Tani, and Tom Walz for identifying the errors and pointing them out to us.”
MacKinnon, of course, shared the 2003 Nobel Prize in Chemistry.
The notice is signed by all of the authors. Kudos to them for swift action.
Well done for fast work.
On another note, a great way to motivate journals would be, for each retraction, a journal lost a percentage of its IF.
It is doable.
You must be careful about incentives here. I assume your suggestion is made with the valiant goal of reducing the number of papers in the journal initially appearing that would be retracted later by providing an incentive for the editors to be more vigilant. I’m afraid that your suggestion may, in all practical cases, merely discourage editors from retracting papers that should be removed. After all mistakes will always be made, and your suggestion would punish editors that correct them, and reward those that do not.
If these mistake were detected so fast by so many, why did they get by the reviewers?
If you look into the paper it was edited by Christopher Miller and this manuscript was submitted to PNAS directly with this pre-arranged editor. Sounds very fishy. Thus, it was a total sloppiness by Chirstopher Miller and he should be expelled from the editorial board of PNAS.
I thought PNAS had done away with that direct submission category?
It was a Direct Submission. Any submitting author may pre-arrange a Monitoring Editor.
David Peterson: What evidence do you have that something “fishy” occurred here?
Its because the reviewers don’t review. This is a growing problem, a very big problem to be precise. I’ve had papers come across my desk where my team and I meticulously tore it apart, finding all sorts of massive errors, misleading interpretations and overstated conclusions. I learned later that there was no feedback from the other reviewers who simply gave it their ‘publicate now’ approval. Science cannot work like this when the peer-review system breaks down. You either find the time to actually review the work or you simply be honest and say that you cannot review at this time. Its astonishing.
Can someone with knowledge in this area explain the error? The wording makes it sounds like a very basic oversight reflecting lack of knowledge by the authors, reviewers, and editor. But that seems unlikely given the prominence of the editor (Chris Miller). Anyone?
Impressive to see correction and retraction work as it should, promptly, with no fuss, and with proper credit. Compare with Michael Mann and friends, for whom stonewalling and obfuscation is the more usual tack:
still no Mann et al. (2008) PNAS retraction
A rare example of honest error..
In this case – like all papers with structural models, the reviewers would not be able to see the raw data (initial images for indexing, final structure factors, etc.), so there is no way to detect the errors indicated by the authors – reviewers just have to trust the images in figures as they are presented. After publication, coordinates and structure factors are released. At this stage, any interested party could look at the model with appropriate electron density maps or initial diffraction images and detect the problems mentioned.