Title: A mitochondrial ferredoxin is essential for biogenesis of cellular iron-sulfur proteins
What Caught Our Attention: Here’s a cut-and-paste issue that gave us pause. The authors of an 18-year-old paper in PNAS corrected it after realizing some bands in a figure were duplicated (an issue raised on PubPeer one year ago). It turns out, the first author had cut the paper into pieces and reassembled them to present the blots in the “desired order,” and some had become duplicated by mistake. The overall results were unaffected, so the journal swapped the image with a corrected version.
Journal: PNAS
Authors: Heike Lange, Anita Kaut, Gyula Kispal, Roland Lill
Affiliations: Institut für Zytobiologie und Zytopathologie der Philipps-Universität Marburg, Germany
The authors wish to note the following: “Recently when we analyzed Fig. 4B (a Western blot visualizing the yeast protein Leu1), we noticed that there was an error in the assembly of the bands. Because the gel loading for the Western blot did not fit the desired order for presenting the data, the first author cut the thermoprinter paper printout into pieces and reassembled them. Unfortunately, the published version contained an unintentional data duplication (mirrored bands for the 64 h data points), but the overall results were not impacted. We apologize for the mistake.” The corrected Fig. 4 and its legend appear below.
Date of Article: February 2000
Times Cited, according to Clarivate Analytics’ Web of Science: 200
Date of Notice: March 12, 2018
Like Retraction Watch? You can make a tax-deductible contribution to support our growth, follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up for an email every time there’s a new post (look for the “follow” button at the lower right part of your screen), or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].
So, what is this hapless “first author” doing these days?
https://www.ncbi.nlm.nih.gov/pubmed/?term=Heike+lange
Oh come on. Mistakes are made, at least I make enough of them for the both of us. The important bit is how you deal with it once you become aware of them.
How does a mix-up with thermoprinter output produce mirrored bands?
https://imgur.com/FNR99ji.jpg
This explanation makes no sense. The bands are not just duplicated (which in itself would have required cutting up more than one copy of the blot–why would you do that?) but MIRRORED as well. Where did the mirrored images come from?
Again, this is Heike Lange speaking. Please see my explanations below. to answer to your specific questions:
– why more than one copy of the blot: I made at least five copies of this figure by gluing excised band in a paper layout – one for the journal, one copy for each reviewer and two for the archive. They were send as paper mail, because that was the usual procedure in 2000.
– How can bands be mirrored: By accidentally photographing the blot upside down or from the wrong side without noticing before cutting the bands. The procedure was not idiot-safe, I’d say today, and I obviously did not pay enough attention.
unintentional data duplication
One lane is triplicated, appearing in the 16 hr, 40 hr and 64 hour sections of the band. The authors should consider making an Erratum to the Erratum.
https://i.imgur.com/FNR99ji.jpg
This is Heike Lange, the first author of this erroneous figure.
I made this figure by cutting out thermoprints because
we did not use photoshop at the time but used a camera similar to todays babyimager connected to a thermoprinter. I had a paper template of the figure layout, but the size and the spacing of the layout did not fit to the prints (the only possibility to adjust the size was to adjust the distance between the camera and the x-Ray). In addition, the experiments had extra control samples that were not mentioned in the text and therefore had no space in the layout.
Not being aware of the possible pitfalls, I made several prints of the figure, cut each band, and clued them in the desired place of the template. Of course this is a procedure designed to make mistakes rather than to avoid mistakes, but I the time I did not realise this. It would have been better to change the layout/the text, or to re-run the gel to fit the layout, but at the time of making I did not think about it. And i did also not go to discuss it with my supervisor, because, as said above, I was not aware of potential problems and just happy that I solved what I perceived as my main problem, which was adjusting the size of the prints.
Some paper copies of the original thermoprint copy-clued layouts remained in the archives, and none of the remaining copies shows the duplicated/mirrored band that is seen in the published figure.
It was a honest mistake that nobody spotted in time.
The originals that do not contain any duplicated (or triplicated band if you think that this really matters in this context) are shown in the corrected figure, and prove that we did not observe any significant changes in Leu1 protein levels in this experiment. Personally, I do not see any need for further actions from my side, and consider the case as closed: i made a mistake, I have no problems to admit it, I corrected it. However, you’re totally free to continue commenting it for the next 18 years.
I think the author here seems to have a misunderstanding on the meaning of the word “original”.
The first claim, is that original gels/blots were run with extra control lanes and in the wrong order. This is what led to the splicing/rearrangement being necessary.
Now in the correction, it is claimed that the corrected figure shows “The originals that do not contain any duplicated bands”. If this corrected image is in-fact an “orginal”, why are the bands all in the correct order?
These 2 statements are factually incompatible. Either there was an original without any extra bands that did not need splicing, and that’s what is presented here in the correction. OR the thing being referred to as “original” in this correction is not original, and still contains a splicing seam.
Given that the corrected image came from “the archives”, and that cellphone cameras are ubiquitous now, it would be a relatively simple matter to snap a picture of the actual physical piece of paper that was glued, under natural lighting (maybe even the lab book page?) in order to verify whether or not the image now being presented as “original” is in-fact so.
Sorry for the misunderstanding. Indeed, I am using the word “original” twice for two different things. To clarify:
The original X-rays contained the experiment with extra bands. I took several photos from it in different magnifications to make the size of the bands fitting to the paper layout and printed them with the connected thermoprinter.
As far as I remember, I cut one piece for each band, at least for the six bands in the right part of the figure. I did mistakes while doing it, resulting in wrong figures with duplicated and mirrored bands destined for the bin. Once I got it right, I took again a picture of the correctly assembled stripe and printed them 5 times (to save time and, ironically, because I thought it would avoid mistakes. By contrast, I can not remember that I was concerned with the splice sites being visible or not) and finally produced at least 5 of these papers one for each copy of the entire manuscript.
Obviously at least one of these papers received a wrong stripe that was accidentally made and should have been gone to the bin. As mentioned before, the procedure is incredible error prone, a clear “how-not-to-do-it”, but at the time of making I did not really realize it.
The best I can say is that I learn from my mistakes.
Two of these old paper figures with the thermoprints were still in the office archive. I referred to them as the original paper figure, but I agree that this is misleading.
Both of the two old paper figures are identical, and one was send to the journal, by physical mail. The experiment was also repeated by somebody not involved at the time, and this new experiment was also provided. I could not provide the original X-rays because I throw them away when I moved into a flat with little storage space about four years ago, assuming that I would never need them again. Also a mistake, but also not intended.
The corrected figure is a scan of the paper figure that was send to the journal. I can not make a photo from it because I do not have it. As it is not a digital remake we left it as it was, but indeed, signaling the splice lanes between the lanes would have been a good idea.
I understood Dr Lange’s explanation differently. See the supplementary comment, located further up in the comment thread:
“– why more than one copy of the blot: I made at least five copies of this figure by gluing excised band in a paper layout – one for the journal, one copy for each reviewer and two for the archive.”
I understood the explanation here — “Some paper copies of the original thermoprint copy-clued layouts remained in the archives, and none of the remaining copies shows the duplicated/mirrored band that is seen in the published figure” — to refer to these copies that were reassembled at the same time as the published version, but without the duplication and mirroring.
It remains entirely unclear to me how the “mirroring” happened, if indeed the questionable image is a true mirror image of the original image. Were some of the copies on (very) translucent paper, which in the event was assembled upside down? Was the copy machine equipped with hardware or software that could make mirror copies by mistake (I’d be surprised if it even had the capacity to make them deliberately, but let’s be generous here)? Was there an intermediate stage in which the paper copies were converted to film negatives and some of them then turned upside down before being used to make new paper copies? ‘Tis a puzzlement.
the x-rays are translucent, and I assume that I put them upside down in the camera box. The camera was directly connected to the printer, no negatives. The printer produced stripes that I cut in pieces. I must have cut one of the upside down stripes into pieces without realizing it.
Try it out, it’s very easy to make mistakes and I am the last person to deny that. If needed, we can organize a meeting on how-better-not-to-do-paper-clipping, but this technique to produce print layouts is quite outdated not to say completely obsolete. I totally agree with the point that I should not have done it? What else can I do?
I think the real point this discussion is: I say it was an unintended mistake, and the best evidence I can provide to support my statement is that we still have paper figures that do not show the duplicated/mirrored bands. I’m perfectly aware that this is not the same as showing the original X-ray, but its the best I can do for now as the X-rays were thrown after ten years of storage.
The journal didn’t take it easy, and I assume that they thought thoroughly about it before they accepted this for correcting the figure.
I can live with having made this mistake, and I guess I will also have to live with people that assume I did it on purpose.
Just for my curiosity, who doubts the conclusion of the experiment, i.e. that mitochondrial ferredoxin is required for assembly of cytosolic iron sulfur proteins?