A paper that shares a first author with a paper retracted in December has been corrected.
Late last year, we reported on a retraction in Antioxidants & Redox Signaling (ARDS) by Indika Edirisinghe, who was at the University of Rochester when the original paper was published, and colleagues. On January 17, the Journal of Agricultural and Food Chemistry published a correction to “Effect of Black Currant Anthocyanins on the Activation of Endothelial Nitric Oxide Synthase (eNOS) in Vitro in Human Endothelial Cells,” on which Edirisinghe is also first author.
His affiliation on that paper, originally published in July 2011, is the Illinois Institute of Technology. Here’s the correction:
Western blots in Figure 5 were distorted in publication. Distortion is likely due to a problem with color/contrasting method that was used to remove background smears of the bands to help better visualize the data in print. The corrected figure gives the data derived from the same experiment without changing contrast/color. The graphs and the figure caption remain without any change. Overall, this correction will not affect the interpretation or conclusions of the manuscript.
Figure 5. Effect of vitamin C depleted black currant juice concentrates on the activation of Akt and eNOS in vitro in HUVECs. Vitamin C in the black currant sample (Ben Hope, BH) was eliminated using the method described under Materials and Methods. Representative immunoblots show the effect of BH and vitamin C depleted BH sample on phosphorylation of Akt (p-Akt) (A) and eNOS (p-eNOS) (B). Vitamin C (respective concentration found in BH sample) alone increased the level of p-Akt and p-eNOS significantly (P < 0.05). However, significant difference was not observed between BH samples and vitamin C depleted BH sample (P > 0.05). The histograms shown in both panels C and D are those obtained after quantification of the blots using densitometry (n = 3) for p-Akt and p-eNOS, respectively. The ordinates are the relative ratios of the phosphorylated and nonphosphorylated forms of each enzyme. (
) P < 0.05 and (
The paper has been cited six times, according to Thomson Scientific’s Web of Knowledge. We’ve contacted Edirisinghe to ask whether there would be any more corrections or retractions, and will update with anything we learn.
While we have you, Retraction Watch readers also uncovered a 2009 retraction of a Molecular Nutrition and Food Research paper and a 2008 correction in Respiratory Research by the senior author of the retracted ARDS paper, Irfan Rahman.
Hat tips: StrongDreams, Fernando Pessoa
Thanks Ivan and Adam! This case deserved more attention, since hopefully several retractions/corrections are coming as eluded to in the comments section regarding the Antioxidants & Redox Signaling paper.
“a problem with color/contrasting method that was used to remove background smears of the bands to help better visualize the data in print”
What’s amazing is that the journal fell for this, when it’s quite obviously blot splicing. The images are in B&W, so what do color problems have to do with it?
Oh, and in Figure 6A, Akt blot (lower panel), the first 3 lanes have been duplicated and rotated 180 degrees to yield the last 3 lanes. No way that’s gonna be ‘splained away by some freaky-deaky contrast method to remove smears!
“Oh, and in Figure 6A,”
What does one do in a situation like this? Email the journal? Just let it go and gripe on a blog?
Email the journal.
Yes, email the journal about this meaty story. I hope the individuals concerned don’t get-a-grilling.
I may go up in flames for this, but, this story will get tastier with age.
In reply to Stewart February 4, 2013 at 5:41 pm
2 papers in the same field of respiratory physiology. A comparison would be useful.
1. J Immunol. 2007 Feb 15;178(4):2491-8.
2. PLoS Med. 2009 May; 6(5): e1000076.
“first 3 lanes have been duplicated and rotated 180 degrees to yield the last 3 lanes”
It isn’t obvious to me.
Besides the phospho-Akt blot above looks reasonably similar to its pair below (a slight angling)
I was under the impression that phospho-proteins were supposed to be probed for on the same strip that you probed for total protein, but no one seems to confirm that point when I raised it recently regarding an American group in La Jolla, so perhaps I am mistaken.
Some reposts, but this group has some issues:
Flip the two uppermost far-right IKKa and p-s176/180-IKKa blots in 6C and you will notice that they are identical:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2396248/figure/fig6/
Control and H2O2 (4 hr Hoehst dye) in Figure 1A are identical. PMID: 15456740 FASEB J. 2004 Dec;18(15):1897-9. Epub 2004 Sep 28. Oxidative stress and cigarette smoke alter chromatin remodeling but differentially regulate NF-kappaB activation and proinflammatory cytokine release in alveolar epithelial cells. Moodie FM, Marwick JA, Anderson CS, Szulakowski P, Biswas SK, Bauter MR, Kilty I, Rahman I.
Any ideas why people do the flipping? One too many at the Friday beer hour?
It is so convincing when a whole row has been flipped. When is one band here, and one band there, and you desribe it as “mirror image” people often don’t believe it.
“Any ideas why people do the flipping? One too many at the Friday beer hour?
It is so convincing when a whole row has been flipped. When is one band here, and one band there, and you desribe it as “mirror image” people often don’t believe it.”
I wish I knew. On the one hand, flipping an image you are already using seems like a surefire way to get caught. Why not get one from another gel? So maybe it’s an accident. On the other hand, the “rotate” and “transform” commands in photoshop and powerpoint aren’t sitting right on the desktop so you can trigger them with an accidental mouse click. And if you are doing so much image manipulation that you can accidentally lose track of an image and reuse it, then you are probably doing too much image manipulation in general.
I kind of raised this in the other post. It is easier for me to understand fabricating a Western that you never ran, than to understand why you would fake up an image of a result you believe in because you are too lazy to re-run the gel.
“I was under the impression that phospho-proteins were supposed to be probed for on the same strip that you probed for total protein, but no one seems to confirm that point when I raised it recently regarding an American group in La Jolla, so perhaps I am mistaken.”
It’s probably best practice, but not by itself enough to raise a fuss over. For one thing, stripping is not always 100% effective. Doing phospho-, then stripping and doing total, works if the total is so much stronger signal than phospho that you don’t get a contribution from the carryover. So if the phospho ab is higher affinity, the sequential method might not work. If we allow that it is acceptable to measure (for example) NF-kB on one blot and MAPK on a different blot, then it’s not much different to run pMAPK on one blot and MAPK on another, as long as you load them identically and have a loading control on both gels to verify uniformity.
That said, if you run different blots, you should definitely not photoshop them to look more similar. Just say in the legend, run on different blots but verified with appropriate loading controls.
In reply to Junk Science February 5, 2013 at 2:02 pm
Is there a field effect?
J Inflamm (Lond). 2010 Jul 16;7:33
http://www.ncbi.nlm.nih.gov/pubmed/20637110
Might be! Nice one, the Fig 2 WT lanes where they don’t even align them properly is a favorite.
Hint: compare figures 1 and 2.
Fig one and 2 are same but they also show the same thing. Probably part of the same experiment.
Figures 1 and 2 are not the same, parts are the same.
The PI3Kgamma-/- Lamin A/C panel in figure 2 is the same as the Lamin A/C panelin figure 1.
That’s image reuse for a start whether it is the same experiment or not.
In the text we are told that in figure 1 the starting material is wt.
Maybe they switched loading controls between wt and mutant in fig2. Also if u notice the graph values of wt in fig1 and 2 are the same. Thats y I said it looks like one exp split into 2 for the paper. Whether they should have used a different blot for the same exp in2 different figures. Maybe. But we should be very careful judging others work.
No, that’s what they have published for people to see.
“Also if u notice the graph values of wt in fig1 and 2 are the same. ” is another issue.
2 papers in the same field of respiratory physiology. A comparison would be useful.
1. J Immunol. 2007 Feb 15;178(4):2491-8.
2. PLoS Med. 2009 May; 6(5): e1000076.
I do not understand “Maybe they switched loading controls between wt and mutant in fig2.”.
I am not saying aurhors are right or wrong. The reason I defended them was I did not agree with the tone of critisim which is fairly common here. What I wanted to convey was that we should respect other peoples woek. If we find it to be wrong (intentional or otherwise) we should repot it to appropriate channels. Innocent until proven guilty.
In reply to tsh February 7, 2013 at 12:36 pm
I was not criticising you, just asking questions. I hope you were not upset.
The paper we were talking about J Inflamm (Lond). 2010 Jul 16;7:33 is not that important.
More important would be to look at:-
1. J Immunol. 2007 Feb 15;178(4):2491-8.
2. PLoS Med. 2009 May; 6(5): e1000076.
There is a point to this. I would like others to comment on any shared things between the papers.
I want to leave it open.
Dear fernando p,
In J Immunol. 2007 Feb 15;178(4):2491-8 Fig. 4, the total p38 MAPK bands in part B lanes 1,2,3,4 look remarkably similar to the total p38 MAPK bands in part C lanes 2,3,4,5, yet the labeling seems to indicate the lysates came from cells receiving different treatments.
In reply to michaelbriggs February 8, 2013 at 12:45 am
Can you find any relationships between figure Figure 1 PLoS Med. 2009 May; 6(5): e1000076.
http://www.ncbi.nlm.nih.gov/core/lw/2.0/html/tileshop_pmc/tileshop_pmc_inline.html?title=Click%20on%20image%20to%20zoom&p=PMC3&id=2674207_pmed.1000076.g001.jpg
and Figure 7. J Immunol. 2007 Feb 15;178(4):2491-8?
http://www.jimmunol.org/content/178/4/2491/F7.large.jpg
OK,
PLoS Med Fig. 1A GATA3 0 min cytoplasm looks like J. Immunol Fig. 7B nuclear 30min,
PLoS Med Fig. 1A GATA3 30min nuclear looks like J. Immunol Fig. 7B nuclear 30min
PLoS Med Fig. 1A histone 1 0 min and 30min look like J. Immunol Fig. 7B nA histone 1 0 min and 30min
In reply to michaelbriggs February 8, 2013 at 11:45 am
That’s bit of a coincidence.
“PLoS Med Fig. 1A GATA3 0 min cytoplasm looks like J. Immunol Fig. 7B nuclear 30min” is problematic.
http://ar.iiarjournals.org/content/33/2/744.full
That’s unrelated to any of these authors and also a good example of proper science, I think. Maybe the controls in the original study were faulty, but admitting that the data can not be reproduced important. I wonder how many other authors publish a paper, can’t reproduce the results for a follow-up, and just quietly drop it.