Twitter is abuzz today over allegations that a recent paper in Scientific Reports contains a blatant example of duplication.
According to the allegations, a group of researchers in Malaysia have used the same four images to represent some 30 cells at different stages of cell death. One researcher has even suggested the allegedly doctored images appear in three different papers.
Is this a manipulated image? See for yourself:
WTF? https://t.co/efHzfpt39q via @quokka79 pic.twitter.com/GlnorC7Wmn
— Christophe Leterrier (@christlet) June 10, 2016
The paper “Novel piperazine core compound induces death in human liver cancer cells: possible pharmacological properties” was published in April.
We contacted first author Nima Samie, affiliated with University of Malaya in Malaysia, who denied that the cells represented the same four images duplicated throughout the figure. Rather, all “1,” “2,” “3,” and “4” images look the same because each represents a cell in the same stage, Samie said:
Of course cells in each step should be same with each others. For instance, all cells which underwent blebbing must be same until they confirm as blebbed cells based on morphology of the cells. This test is 100% based on morphology of cells only…They can not be copy pasted cells but morphology is same.
We asked why the cells look exactly the same. He said:
With respect to others opinion but of course they should be same because they are in same stage of apoptosis. If different, they can not be considered as same category. We tried to capture the best images until most of the cells in each stage be same to each other so we could show at least enough number of cells are present to be reported. not only 1 or 2 cells.
But others aren’t convinced. Another researcher suggested that the duplication extends across multiple papers:
https://twitter.com/Alexis_Verger/status/741238641327017985
Those other two articles are: “Mechanism of Action of the Novel Nickel(II) Complex in Simultaneous Reactivation of the Apoptotic Signaling Networks Against Human Colon Cancer Cells” published in frontiers in Pharmacology in January. “Mechanism of action of novel piperazine containing a toxicant against human liver cancer cells” appeared in PeerJ in March.
None of the papers have been cited yet, according to Thomson Reuters Web of Science.
A spokesperson for Scientific Reports told us:
We do not comment on papers that may or may not be under consideration for retraction or correction.
All three papers share a first author, and three co-authors. Like first author Samie, co-authors Sekaran Muniandy, Batoul Sadat Haerian, and M. S. Kanthimathi are all listed at the University of Malaya.
Update, June 10th, 12:00 PM EST:
PeerJ has placed an expression of concern on the paper it published:
Questions have been raised about some aspects of the figures in this manuscript. This expression of concern will remain in place while PeerJ staff investigate this matter further.
We asked Elisabeth Bik, a microbiologist at Stanford, and author of a study on instances of image duplication, to weigh in on the Scientific Reports image. She told us:
Yes, obvious copy / paste job. All red cells (#4) look the same. Several #1 cells look the same. Many more problems with other figures in these papers – almost every figure appears to have duplications, manipulations. FACS plots duplicated, Western blot with many splices and duplicated controls, rotated / manipulated panel in Figure 1 (normal cell = SNU-423 and also = SNU-475 with removed cell).
She noted that the paper itself seems to be duplicated:
This paper should not have passed peer review – three times.
Update, June 10th, 1:45 PM EST
On his blog at Science Translational Medicine, Derek Lowe points out other problems with the paper, and writes:
Enough. The authors should be ashamed of sending a manuscript out like this, because its mistakes are so numerous and so obvious as to make outright fraud – deliriously incompetent fraud – the first explanation that comes to mind. The reviewers, if there were any, should also be ashamed for letting something like this pass. Who looked at this stuff?
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.
Thank you for this post. PeerJ was alerted to this situation today and we have placed an Expression of Concern on the article and are investigating.
Next time, follow up with “Why do all the cells have exactly the same orientation?”
Nice to see that in some of the other papers, the effort was at least made to perform 90-degree rotations.
Kudos for good detective work to Christophe Leterrier. The responses and rationale of the first author Nima Samie are preposterous. The probability that two cells with identical blebbing pattern could be found in the culture would be exorbitant. The author’s claim that image in Fig. 6 showing 9 cells with identical blebbing pattern in the same microscope field is real shows simply put ignorance of science.
I must say that I did not do any detective work! Saw a tweet from @quokka79 and checked the paper with incredulity. According to Altmetrics I think the chemists (as stated in Derek Lowe’s blog post) were on it first related to the chemistry described in the Methods.
One of the things that’s the most grimly enjoyable about this series of papers is how they seem to have crudely faked it in every discipline they touch. . .
Is it just me, or the entire paragraph was copied and pasted?
1) In Mechanism of Action of the Novel Nickel(II) Complex in Simultaneous Reactivation of the Apoptotic Signaling Networks Against Human Colon Cancer Cells:
Figure caption/description:
“Supplementary Figure 3. AO/PI double-staining.
Figure reveals that untreated cells remained healthy after 72 h. Moreover, early apoptotic features such as chromatin condensation and blebbing were witnessed after 24 and 48 h. However, after 72 h of NTC treatment (6.0 μM), late apoptosis event was observed. A, Blebbing of cell membrane; B, Late apoptosis; C, Chromatin condensation; D, Viable cells.”
2) In Mechanism of action of novel piperazine containing a toxicant against human liver cancer cells:
“PCC induced membrane blebbing and nuclear condensation
After 24 h treatment of SNU-475, and SNU-423 cells with 6.25 µg/ml of PCC, the apoptotic features were analysed by using fluorescence microscopy. Normal nuclear structure in control cells was displayed as green fluorescence, whereas bright green fluorescence was shown in early apoptotic cells, caused by interposition of acridine orange with the fragmented DNA. Nuclear chromatin pucker and membrane zeiosis, as moderate apoptotic features, were detected after 24 and 48 h. Binding of propidium iodine to denatured DNA was identified by reddish-orange colour after 72 h indicating the late stage of apoptosis (Fig. 7).”
Figure caption:
“Figure 7: AO/PI double-staining.
Early apoptotic features such as chromatin condensation and blebbing were witnessed after 24 and 48 h of PCC (6.25 µg/ml) treatment. However, after 72 h of PCC treatment (6.25 µg/ml), late apoptosis event was observed. (A) Blebbing of cell membrane after 48 h post-treatment. (B) Late apoptosis after 72 h post-treatment. (C) Chromatin condensation after 48 h post-treatment. (D) Viable cells after 72 h post-treatment.”
3) In Novel piperazine core compound induces death in human liver cancer cells: possible pharmacological properties:
“PCC induced membrane blebbing and nuclear condensation.
After 24h treatment of SNU-475, and SNU-423 cells with 6.25μg/ml of PCC, the apoptotic features were analyzed by fluorescence microscopy. Normal nuclear structure in control cells was displayed as green fluorescence, where as bright green fluorescence was shown in early apoptotic cells, caused by interposition of acridine orange with the fragmented DNA. Nuclear chromatin pucker and membrane zeiosis, as moderate apoptotic features, were detected after 24 and 48h. Binding of propidium iodine to denatured DNA was identified by reddish-orange color after 72h indicating the late stage of apoptosis (Fig. 6).”
Figure caption:
“Figure 6. AO/PI double-staining.
The Figure reveals that untreated cells remained healthy after 72h. Moreover, early apoptotic features such as chromatin condensation and blebbing events were witnessed after 24
and 48h. However, after 72h of PCC treatment (6.25μg/ml), late apoptosis event was observed. 1: Viable cells; 2: Blebbing of cell membrane; 3: Chromatin condensation; 4: Late apoptosis.”
All three papers should be retracted immediately – and then republished, in a textbook as an example of dodgy data.
It’s ridiculous. Even if only one of the journals published this article, I can’t believe the reviewers would miss blatant photoshopping. For those of you who’ve missed it, check out Fig. 2 in the Frontiers paper. The nuclei in ‘normal’ cell image are IDENTICAL to the HT-29 (with a couple deleted), and just rotated in WiDr!
…and it looks like they’ve manipulated the westerns as well… the b-actin is identical in all conditions and it appears other lanes have just been copied over.
And the journal. I recently had the experience of having a journal come back and query an image of mine – the typesetters (or whatever the modern equivalent are) had looked at the image and decided it was over exposed and queried the result. They were absolutely right to do so, it was one of the first images I ever took on a deconvolved microscope, I didn’t really know what I was doing at the time and it was indeed over-exposed. As I prepare manuscript images as I go along, it had just stayed in the paper without me really questioning it. I repeated the experiment, redid the image, submitted and all was well with the world. However, clearly this journal didn’t bother doing any basic checks on the images as surely the photoshopping would have been revealed.
The story has also been covered by Dr. Leonid Schneider:
https://forbetterscience.wordpress.com/2016/06/10/triplicated-paper-with-multiplicated-cells-and-images/
published in frontiers in Pharmacology
These people have some explaining to do.
i agree.
esp since he’s “director of research” (http://loop.frontiersin.org/people/36464/overview)
The journal site only gives reviewers names and not their recommendation. Editors need to clarify whether they overrode reviewer’s recommendation. A rejected paper is lost revenue in the OA journal.
Correct. By all accounts, it is quite difficult for a reviewer to have a paper be rejected in Frontiers and also for the reviewer to have their name disassociated from the review if the editor overrules their recommendation.
There is absolutely no point in making reviewers’ names public unless their reviews/recommendations are also public.
The case is reminiscent of the “Sensor and Actuators” paper including a grotesque TEM image:
https://pubpeer.com/publications/CD224AAA5BE156801AB113332C6B0B
More than a year after the first Pubpeer alert, nothing has happened. No retraction, no correction, not even a discount on purchasing this article. And the author for correspondence remains in place in the editorial board of the journal…
The link for the google.scholar page of Nima Samie (1st author) is now dead.
But he still has a LinkedIn page: https://www.linkedin.com/in/nima-samie-60866860
Universally, the common ‘rhetorical’ question around the table when ORI discussed an image case was “who reviewed this turkey”? That is not to necessarily blame journal reviewers, nor coauthors for that matter, but to point out how it still remains astonishing that stuff like this is not originally seen (i.e., ‘perceived’) by those who were ‘experts’ in the research subject. Julio Turrens once reported in an abstract at an 2012 ORI conference that he and his students identified many more image problems in papers when they looked at the images before reading the paper. The theory was that then the reviewers perceptions of the data would not be blinded by the supposed beauty of its conclusions. Try it when you review the next paper! How much of research misconduct arising in publications can be prevented by small changes in the ‘practice of conducting science’ rather than grand, exhortative lamentation about behavior and its standards?
Several other papers by the corresponding author have image problems too, such as DOI: 10.1038/srep25139 (Figure 3), DOI: 10.3109/19396368.2015.1112699 (Figures 5 and 6), and DOI: 10.1016/j.theriogenology.2015.09.036 (Figure 1).
Posted a thorough analysis of the paper, nearly every figure is manipulated.
How did no one catch this?
https://pubpeer.com/publications/F30641CE24023236C0C54A9229F69E#fb52589
The ResearchGate profile of Nina Samie has disappeared:
http://www.researchgate.net/profile/Nima_Samie
What does the institute, the University of Malaya, have to say?
I can still find and go to that page when logged in and searching for his name.
“His page”? Is Nina a man?
Yes.
Nina Samie may not be, but his first name is Nima.
Oops! My spelling mistake! Apologies to Nima. This still does not change the fact that without logging into ResearchGate, the link to the profile above is not visible to the public, but I am no sure why.
There is a total of 4 papers from the same first author that are very similar, both by text as well as by figures. I have made a side-by-side comparison and show more examples of the possibly duplicated figure panels here: https://microbiomedigest.com/2016/06/10/image-duplication-quartet/
well you only need to look at the bad job that he’s done with microsoft paint: http://imgur.com/Z0WBYLs
Someone posted the following image on the Derek’s blog:
http://imgur.com/Z0WBYLs
Look carefully at the background for cell (1) on the right…
This is just ridiculous and depressing.
It would be bad enough if this was one instance in an obscure pay-to-pub journal, but three including a NPG journal?!
Editors? Peer reviewers? Anyone??
I am aghast… How could scientific reports do this ? I am told they expect the authors to upload images of every experiment. If a Western is done in triplicates, images of all the three replicates should be uploaded as supplementary text.. If this is true, HOW could this not be detected…
Because at the end of the day , what matters is money – the so-called “cheque-book publishing”.
The problem is that Nature Publishing Group has fused to Springer Science + Business Media to form SpringerNature. Can we expect the same quality control there?
I now get an internal server error (500) when I got to the SciReports link: http://www.nature.com/articles/srep24172
It’s working again. I think they had a general problem with the homepage, as I could not get access to any papers yesterday.
Universiti Malaya (University of Malaya) is ranked 27th in Asia:
http://www.topuniversities.com/university-rankings/asian-university-rankings/2016
Look closely, results for SNU-475 is totally same as SNU-423, the author manipulated the image (just rotate the image, and adjust the brightness).
University Malaya Faculty of Medicine under investigation:
http://www.thestar.com.my/news/nation/2016/06/15/idris-to-probe-falsified-research-data-claims-at-um/
Refered to (Figure 2. NTC arrests the cell cycle in the S/M phase)
Again, image manipulation. Normal cell is same as HT29
It is sad to see that the only statement from the Journals comes from PeerJ. One would assume high impact NPG journals handle such issue somewhat more professional.
Also that the SciRep servers were down multiple days and not a single statement during or after this incident from Nature.com does not help in their reputation.
Latest update:
http://www.thestar.com.my/news/nation/2016/06/16/committee-finds-duplication-and-manipulation-of-research-data/
That was fast! I guess it wasn’t all that difficult to determine the blatant duplication, but still…the committee is done already with its work.
http://m.malaysiakini.com/news/346240