We have a follow-up from last week’s story about a University College London (UCL) investigation into the work of Giulio Cossu that found errors but no “deliberate intention to mislead.”UCL said it will not make the full report available:
…on the advice of UCL’s Freedom of Information officer UCL will not be releasing the report given that it includes information provided in confidence, as well as personal information.
The university’s panel looked at eight of Cossu’s papers highlighted by pseudonymous whistleblower Clare Francis. One will be corrected: “Magic-Factor 1, a Partial Agonist of Met, Induces Muscle Hypertrophy by Protecting Myogenic Progenitors from Apoptosis,” which appeared in PLoS One in 2008. The paper has been cited nine times, according to Thomson Scientific’s Web of Knowledge.
UCL also told us that they “did not consider” a now-retracted Journal of Cell Biology paper as part of its investigation, “but simply noted it.” They sent the relevant paragraph from the report:
In his communications with the Panel, Professor Cossu had highlighted an article published in theJournal of Cell Biology that had subsequently been retracted. He explained that this was not from his group but from a neighbouring laboratory that included him as a senior author on the basis of his mentorship activities (J. Cell Biol.172, 233-44, 2006). The Chair explained to Professor Cossu that the Panel had noted this article, but wished only to consider the specific allegations submitted by ‘Clare Francis’, in line with the Panel’s terms of reference.
So perhaps
“Because the queries were made anonymously, it was considered impracticable to pursue them through UCL’s usual procedure for investigating allegations of misconduct in academic research.” (from last week’s post)
equates with
“The Chair explained to Professor Cossu that the Panel had noted this article, but wished only to consider the specific allegations submitted by ‘Clare Francis’, in line with the Panel’s terms of reference.”
I’m not sure what to think of that…
Watch “Yes Minister”. You only need one episode to understand!
“He explained that this was not from his group but from a neighbouring laboratory that included him as a senior author on the basis of his mentorship activities.”
Smells like another line, that of “honorary authorship,” got crossed here. Your name is on it, you own it.
Agreed, but of course UCL would say ‘We are not investigating the issue of honorary authorship here, only the specific allegations submitted by ‘Clare Francis’, in line with the Panel’s terms of reference….’
UK universities have perfectly adequate policies on investigating research misconduct (on paper) but the little evidence available (most enquiries never see the light of day) indicates that in the end they just come to the conclusion they want. The only exception was the case at St. George’s Hospital where Malcolm Pearce faked a whole study – the main authors had to fall on their swords. Unless someone can prove that data have been fabricated in their entirety, these committees will always attribute image manipulation to ‘sloppiness’ and describe it as an ‘error’.
“UK universities have perfectly adequate policies on investigating research misconduct (on paper) but the little evidence available (most enquiries never see the light of day) indicates that in the end they just come to the conclusion they want.”
YES, this is the case with most universities – they declare that they will Do_The_Right_Thing in cases of misconduct, when in fact they Do_Anything_To_Cover_It_Up.
Transparency Index would show whether the institutions Did_The_Right_Thing when evidence for misconduct is presented or, on the contrary, they Did_Anything_To_Cover_It_Up.
IT’S TIME FOR A CHANGE!!!
exactly.
I bet that the correction will only address the issues in figure 5 and ignore the other issues.
Figure 5.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2528937/figure/pone-0003223-g005/
Figure 5A. WT panels. The architecture of the upper 3 d panel looks very like the lower 7 d panel.
‘The Chair explained to Professor Cossu…’
You would have thought that Cossu would be the one to do the explaining, or at least providing some information to the investigation about what happened. For example, given the level of fraud in the JCB paper, is the neighbouring laboratory’s institution going to undertake an investigation? That’s a question deserving of an answer given that Cossu was an author.
Anyway, it will be interesting to see what the Correction says when it finally appears; specifically does it involve image manipulation? The implications of that would be significant as it would suggest the problem discussed in the other thread – institutions choosing to redefine image fraud as ‘sloppiness’.
I did inform Tim Perry, the person dealing with the complaints about the work of G Cossu, about the retraction in JBC.
Please see the e-mail list below.
———- Forwarded message ———-
From: clare francis
Date: Tue, Jan 22, 2013 at 6:48 PM
Subject: retraction of prof G Cossu’s work Published January 22, 2013: more specific concerns J Cell Biol. 2006 Jan 16;172(2):233-44.
To: [email protected]
Cc: [email protected], [email protected], Parker Rosa , [email protected], James Parry , fgodlee , cope_opsmanager , VC , “Perry, Tim” , [email protected], [email protected]
———- Forwarded message ———-
From: Liz Williams
Date: Tue, Jan 22, 2013 at 6:25 PM
Subject: RE: more specific concerns J Cell Biol. 2006 Jan 16;172(2):233-44.
To: clare francis
Dear Dr. Francis,
The paper from Pisconti et al (2006), about which you raised concerns, has been retracted from JCB. The retraction has been posted online at http://jcb.rupress.org/content/early/2013/01/21/jcb.2005070832003r.
Sincerely,
Liz
Liz Williams, PhD
Executive Editor, The Journal of Cell Biology
The Rockefeller University Press
1114 First Ave, Floor 3
New York, NY 10065
Phone: 212-327-8011
Fax: 212-327-8576
Email: [email protected]
Website: http://www.jcb.org
From: clare francis [mailto:[email protected]]
Sent: Monday, November 05, 2012 5:48 PM
To: Liz Williams
Cc: Pamela Carpentier
Subject: more specific concerns J Cell Biol. 2006 Jan 16;172(2):233-44.
J Cell Biol. 2006 Jan 16;172(2):233-44.
Figure 4.
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=2063553_jcb1720233f04.jpg
Figure 4A. Satellite cells set of panels (left set of panels).
Please see the Follistatin panel, band in the 2nd lane (starting from left).
This band has a spot beneath it.
Please see the Myostatin panel, the very next panel down, band in the left-most lane.
This band has a spot beneath it. It is in the same relative positions to both bands.
Note that both bands have an area of brightness near their right ends, a little bit closer to theier tops than bottoms. Note that both bands have a band of lesser intensity to their right.
The graininess is different in the two panels. Not all details will be reproduced.
Note that in the Follstatin panel the position of where a band would be in the 4th lane is smooth.
The equivalent position to this is the 3th lane of the Myostatin panel (the next panel down).
There is a band in this lane. Why do you think the position in the 4th lane of the Follistatin panel is smooth?
On Thu, Oct 25, 2012 at 6:00 PM, clare francis wrote:
concerns J Cell Biol. 2006 Jan 16;172(2):233-44.
Figure 1.
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=2063553_jcb1720233f01.jpg
Figure 1E. Error bars small and similar. I know that they are s.e.m.s, but there is little variation.
Figure 1G. Doubt that the GAPDH panel is the loading control for the pCNA panel.
Figure 1H. 4 of the 6 error bars are very similar. One is even smaller, one is larger than the 4 similar ones.
Figure 2.
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=2063553_jcb1720233f02.jpg
Figure 2A. Error bars small and idealised.
Figure 2D. 6 of the 8 error bars are very similar the other two are smaller. There is little variation.
Figure 4.
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=2063553_jcb1720233f04.jpg
Figure 4A. Bands in many of the panels have similar shapes.
Figure 5.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2063553/figure/fig5/
Figure 5D. The error bars are small and do not show much variation.
With respect Clare Francis (and Fernando Pessoa), at times you seem to barking at shadows.
There must have been issues because 1 paper has been corrected and another has been retracted, but far too often I have be clicking on figure links and scratching my head.
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=2063553_jcb1720233f04.jpg
“Figure 4A. Bands in many of the panels have similar shapes.”?? Seriously??
You may have good reason for suspecting the work of this scientist and believing that his work may not be replicable – by which I assume tips from insiders or the like. But unless this malpractice is expressed in the form of photoshop then it is pointless to go a scatter gun approach of raising as much doubts about every single image that you can. In the end you only damage your own credibility.
Unfortunately only a minority of fraudsters are going to be brazen enough to use photoshop and if they don’t use photoshop then it is pointless to pretend that they are. Much as it might be comforting to think otherwise, a group that produces unreplicable work looks very much like a group that produces replicable results.
That’s the danger with vigilantism.
Writing to a journal is not vigilantism, but called writing to a journal.
You may very well be correct, but what we can see is the small window there is.
That others may be more sophisticated is a concern, but not a reason for nihilism and doing nothing.
How many retractions have you managed?
“How many retractions have you managed?” . With all due respect, you make it sound a bit like a bounty hunting contest!
With all due respect, I did not mean to offend your sensibilities.
Why do people have to be wallflowers?
Reply to little grey rabbit May 13, 2013 at 12:08 pm
You don’t make a diagnosis on one thing. You do need to go through each figure pointing out what seems irregular. One thing will not clinch it. It may not be obvious in one figure,or in one paper even, although often it is. I did it by looking at the figures. You don’t need inside tips.
These authors were brazen enough.
Why keep pointing out the obvious that some may be more sophisticated, and that you know others who have always been there and done that?
In reply to little grey rabbit May 13, 2013 at 12:08 pm
You will see that I refined my criticism of figure 4 on Monday, November 05, 2012 5:48 PM
I sent one message on Thu, Oct 25, 2012 at 6:00 PM and then a second message on Monday, November 05, 2012 5:48 PM. The later message is higher up the page. Sometimes other thoughts do come.
I included both, because I sent both.
In any event you do not make a diagnosis on one thing.
There were no inside tips. You can see it on the page.
I think you are quite right that others will hide it better, but what can you do about that?
Thanks for the sentence “a group that produces unreplicable work looks very much like a group that produces replicable results.”, I will be using it a lot (with proper citation of the source!).
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=2063553_jcb1720233f04.jpg
Figure 4A. In satellite cells in the Myostatin and Follistatin lanes I can see the same spots and this “bluring” in “L-name x follistatin”.
There is one more spot in “L-name x follistatin” appearing in “Delta-no x myostatin” (right lower corner).
Overlay (merge) with ImageJ gives a match. This bands are looking suspicious indeed.
On more artifact appears in “E12.5 NT x follistatin” upper right corner and Satellite cells “L-name x myostatin” upper right corner. Also look at this little spot in the left lower corner, both have it (in one it is almost cut out).Overlay with ImageJ gives a match again. Looks suspicious indeed.
The question is: Are those spots artifacts within the gel? Or could they be in the detection system or camera and thereby reappear in different images? If formeris the case, this definitely indicates reused and faked data in this image.
The paper has been retracted. It was not my decision.
Dr Francis,
As (a) you are obviously unhappy about the outcome of the investigation, and (b) many of us consider that you are likely to be more trustworthy and reliable than an institutional investigation of itself, and (c) we can look at the figures ourselves anyway, ..
What do you suggest we do to raise these concerns? I would be happy to send further messages/posts if I thought that others were doing the same.
Michael
In reply to Michael Kovari May 15, 2013 at 11:07 am
Dear Michael,
I think that UCL could have done a better job of it. There seem to be many things about the work which are problematic. I did let UCL know about the retraction in J Cell Biol as soon as the journal let me know.
I do wonder if this is going to be a case of de-discovery.
My message Clare Francis May 13, 2013 at 1:14 pm was a statement. Some seemed to be acussing me of vigilantism so I thought I should make it clear that I had no part in the decision to retract.
My message came out underneath the message of Hans Müller May 13, 2013 at 1:04 pm, but was not a criticism of his looking at figures. I think he does a good job with his comments, is open-minded and keeps to the scientifc points, and I Iike the way he deals with deflections that have nothing to do with the points being discussed.
Journals such as Nature are sold at the newsagents. Nature started out as a journal for the public to read.
I don’t think there was a time when that was forbidden. It is general science journal. Anybody can write to a journal editor.
I suggest that you simply write down what you think looks abnormal about figures, with some reasons, and write a short e-mail to the editors of the journal. If the use of images offends the rules of publication, for example, reuse of images without proper citation, or if the use of images goes against logic, such as using the same images to represent two different things, then you should mention those points. If you think that there is sleight-of-hand you should mention it.
These are not odd things to do.
I do not know if you are familiar with COPE. http://publicationethics.org/
It has flowcharts for how editors should deal with complaints by readers.
Here is the flowchart for dealing with suspicions of fabricated data brought to its attention by readers.
http://publicationethics.org/files/u7140/Flowchart%20Fabricated%20B%20revised.pdf
It is all orthodox. It is not being a vigilante to write to the editor.
Pubpeer is worth writing to. It is actually important to have a centralized system which others can access, and where there may be help and advice.
http://pubpeer.com/
[email protected]
C
I wonder why Nature does not join COPE…
See? At least Claire agrees with me on Pubpeer. The only downpoint in it is that one cannot officially leave a name, thus one will loose the prestige to THE “initiator” of the investigation. Meaning Claire/Fernando/etc having to adbicate of their aliases while complaining. However even this looks a positive aspect to me, as adds sanity to the process and reduces any impression of vigilantism. Pubpeer rocks!
To stimulate controversy here, I tend to agree with littlegreyrabbit. Not all fraud is photoshop fraud. A fraudulent paper can look exactly like a legitimate one, and a group that produces fraud can look exactly like a group that is righteous. I still have my questions about a certain Bohemian monk.
I tend to agree with you. It could be even worse. Just imagine the case of a whistleblower who is actually involved in fraud himself/herself. The perfect cover! It won’t be long before such a case emerges, in my opinion.
You’ve said that before. You can keep taking to the next level ad infinitum.
OK, but does having said it before make it less likely in the real world?
You are quite right, and?
Yes, I agree with many who say that simple similarity in shape is not enough. It reminds me of the issue of cheating on tests. Let us say that one person is sitting next to another person and both are filling in a bubble sheet. Let us say that person 2 has a different test form from Person 1, but has the same identicle pattern of bubbles on the sheet, even though most are wrong.
Would you accuse Person 2 of cheating? There is a strongly suspicious pattern.
I certainly would not.
In the case of a similarity of shape, certainly there must be some tool available which can do a pixel by pixel compare, and render a correlative or summed difference result. However, even a correlation of 1, or a summed difference of 0 or non-zero is still only circumstantial.
Nothing beats the human eye, IMO, for identifying image fraud. I agree some of Fernando Pessoa and Clare Francis throw out there are in a grey area, but others are black-and-white…for example this post from the May 9 thread: http://www.retractionwatch.com/2013/05/09/ucl-finds-errors-in-work-by-biologist-cossu-but-no-deliberate-intention-to-mislead/#comment-56398
A computer algorithm is not needed to quantify the image manipulation in this case.
Yes, the human eye (and ear) is good (those of many animals are even better, by the way). However, the same human eye sees Jesus’s face on a toast and hears voices from the dead in tape recorder noise.
This generalizations don’t make any sense here. Please bring some examples where you saw “the danger with vigilantism” in science. The example Freeloader brought up is very good. You can see the copy-paste and transformation of lanes in Fig. 1A with the eye. Try a flip and overlay with ImageJ of both lanes and then it is proven, that bot lanes are the same. There are many well accepted ways to analyze suspicous data sets (see ORI tools). This should be not that hard to understand…
Reply to Freeloader May 13, 2013 at 2:07 pm
I think that the comment about Hum Mol Genet. 2000 Jul 22;9(12):1843-52.
in this post is near black and white.
http://www.retractionwatch.com/2013/05/09/ucl-finds-errors-in-work-by-biologist-cossu-but-no-deliberate-intention-to-mislead/#comment-56384
I also think that the explanation and significance given in this post is near black and white.
http://www.retractionwatch.com/2013/05/09/ucl-finds-errors-in-work-by-biologist-cossu-but-no-deliberate-intention-to-mislead/#comment-56450
‘Clare’: were UCL aware of the issue with Figure 1 of the 2009 Stem Cells paper of which Cossu is senior author (see above)?
In reply to amw May 13, 2013 at 7:06 pm
Yes. UCL was fully aware from the beginning of the issue with figure 1 Stem Cells. 2009 Jan;27(1):157-64.
It should be “issues” rather than “issue” with Figure 1 of Stem Cells. 2009 Jan;27(1):157-64. In Fig. 1A Pax3 and b-tubulin bands look like mirror images, and in Fig. 1C the null c-kit and Flk1 panels look the same.
In reply to Freeloader May 13, 2013 at 2:07 pm
You are quite right about the grey areas. If there were only grey areas that would be an end to it.
When you find areas of black and white areas it does cast the grey areas in a different light. It is a Bayesian world.
In reply to michaelhbriggs May 13, 2013 at 8:25 pm
Issues about figure 1 Stem Cells. 2009 Jan;27(1):157-64.
which were acknowledged by UCL on Fri, Nov 16, 2012 at 2:42 PM.
http://onlinelibrary.wiley.com/doi/10.1634/stemcells.2008-0503/pdf
Figure 1A.
The black bands in the Pax3 and beta-tubulin panels look like they are resting on a grey band (like the mark in the bath if you do not scrub it).
It simply does not look right.
SUSPECTED DATA FABRICATION (making it up in plain English. Fabrication and manipulation seem to be subdivisions of the same thing).
I believe that the bands in the D16, D15 and D63 lanes in the Pax3 panel are horizontal mirror images of the bands in the D63, D15 and D16 lanes of the beta-tubulin panel. The same bands cannot represent two things at once.
Figure 1B. There is no background.
Figure 1C. Many consider it scientific misconduct to show small strips of blots which are hard to identify (can be more easily passed off as new),
and which do not allow viewers to test them against the laws of physics as there is so little to go on (denying physics).
I think that CD34 null and CD34 wt are different exposures of the same thing.
I suspect that all the images in the C+ column are slightly different versions of the same thing.
I suspect that c-kit null and Flk-1 null are slightly repositioned images of the same thing.
I suspect that Sca-1 null and wt are the same thing.
I suspect that CD31null and wt are the same thing.
I suspect that VE-Cadnull and wt are different exposures of the same thing.
No question Fig. 1A and 1C of this paper appear to have been generated by fraudulent means.
The image manipulation looks trivial and almost childish but we need to remember that it hides deeper problems. For example, WHY did the authors use a mirror image of the Pax3 blot to represent the beta-tubulin expression in these cells (Fig. 1A)? The only answer that makes sense is that they did not do the beta-tubulin blot. And if they did not do it, this is fabrication and misleading to reviewers and readers. And if they are fabricating the western blot, what is the evidence that the rest of the results are real or also fabricated? These are the sorts of questions an investigational panel should be asking.
Did UCL ignore this issue? If so, is it not the appropriate course of action now to inform the journal and at least give them the opportunity to comment? The Editor will probably say that it was up to the university to deal with the matter, but it is still worth trying if anyone has the stomach for it.
In reply to amw May 13, 2013 at 9:10 pm
Editors acknowledged concerns about Stem Cells. 2009 Jan;27(1):157-64
on Fri, Nov 16, 2012 at 8:13 PM
AMW
The mirror image is not easy to see on the untrained eye. Lets spell it out clearly.
http://onlinelibrary.wiley.com/doi/10.1634/stemcells.2008-0503/pdf
Fig 1
1. Look at the first band in fig1A, Pax 3 panel (day 16). Flip it horizontally (left to right) 180 degrees.
2. Look at the 3rd band in fig1A, B-tubulin panel (Day 63).
Now, superimpose the two bands.
You will now see they are identical mirror images of one another.
I don’t know about you, but I miss the big red arrows that make this all too clear. Perhaps we don’t have too much to wait on that front?
If you observe the other bands, try doing the same thing and see which other bands are mirror images of other bands – and this is only one figure in one paper that got through peer review and an investigation.
.
In reply to amw May 13, 2013 at 9:10 pm
http://onlinelibrary.wiley.com/doi/10.1634/stemcells.2008-0503/pdf
Strictly speaking:
Figure 1 is one of the worst figures I have seen so far. This figure is like a slap in the face. Editors are in part aware now. But do all participating authors know about this? Institutes from paris and milan are involved. Do they know? At the end editors have to make the decision about a correction, but co-authors can put some pressure. No reaction since November 16, 2012 is far too reluctant.
Well, you’ve got a point there. Specifically, in fig 1A, the three spots on the left side of the top and bottom rows are mirror images of one another; even I can see that. In fig 1C, there is a remarkable similarity of null and wt spots in the Sca-1 and CD31 lanes. It is not so remarkable in the CD34 lane, but it is conceivable that by adjusting the intensity the two could become much more similar. The other lanes are not so obvious, but given the obvious similarities, it would be worthwhile to play with the intensity and so on in Photoshop to see if the other spots could be made to look the same. Even if they’re not the same, the rest of the figure is compromised.
What does this mean? Prima facie evidence of “image manipulation” if you can match up the images the way they look to be matchable. Why would someone do that? Because they don’t have a legitimate image that shows what they want to show. Why not? They lost it, or maybe they never had it.
I think maybe some of the things Clare Francis says in her/his detailed dissections of images may be questionable, but given the obvious things wrong with these images, there is little room for doubt anyway.
But Ivan’s update here (One will be corrected…) suggests UCL do not recommend correction of this paper.
Or are there more corrections planned? Why is there no clarity?
In reply to amw May 13, 2013 at 10:42 pm
I do not know if any more corrections are planned. The UCL statement reads “Where specific errors were detected, the panel recommended that, where possible, corrections should be made to the published papers”. There has been one correction in Nature http://www.nature.com/nature/journal/vaop/ncurrent/full/nature11976.html
and now we hear in this post about a correction to come in PLOS ONE. Two corrections by themsleves would satisfy corrections so that might be it.
So why is there no clarity? UCL is a business.
Understood – and no surprise. The impression I get is that UCL have broadly tried to minimise the issues, but calculated that the compromises being described were necessary to avoid them being accused of a complete cover-up.
The phrase ‘The panel found no substance to the large majority of these allegations…’ is a special pleading (the word ‘majority’ absolutely does not need to be preceded by the word ‘large’) and placed very prominently in their comment. This is inappropriate in my view because it was Cossu who first published misleading work, and scientific readers need to know first and foremost what was wrong in his papers on medical science. Whether the panel consider , ‘Clare Francis” allegations to be mostly correct or mostly wrong, they were in general terms clearly justified as events have borne out, and we just don’t need the panel to tell us what proportion of allegations had substance. If there were 8 papers, and two have been corrected and one retracted, this is a bare majority. It makes UCL look like dodgy defence lawyers, not a scientific institution.
Out of interest, in the course of these investigations, was there ever any note of thanks for bringing these issues to the attention of the university or the scientific community? Any indication at all that this was an appropriate thing to have done?
In reply to amw May 14, 2013 at 6:38 am
I think that there was haughtiness and obfuscation from the outset.
On Fri, Nov 16, 2012 at 2:42 PM, Perry, Tim wrote:
Dear Clare Francis
I am writing in response to your email below. This has been drawn to my attention as the UCL officer responsible for our procedure for investigating and resolving allegations of misconduct in academic research (available at http://www.ucl.ac.uk/ras/acs/resgov/research-misconduct-procedure.pdf).
As I understand it, the research in question was conducted before Professor Giulio Cossu commenced employment with UCL (in January 2012). Nevertheless, as you will see from the procedure, its scope allows us to investigate research conducted by a UCL employee before or after s/he was employed by UCL. You will see too that the procedure expects an allegation of research misconduct to be made by a named initiator (paragraph 1.5) and (from paragraph 5.1) that ‘The initiator shall be required to produce a detailed statement in writing in support of the allegation and submit any supporting evidence that is available to them. The initiator shall also be required to specify what type(s) of misconduct s/he claims has occurred, with reference to the definitions at section 1 of this procedure’.
Please confirm to me by no later than 6pm on 21 November that you wish me to activate the research misconduct procedure and that, as the initiator of the allegation made, you will confirm your identity and provide me with the information required under paragraph 5.1 of the procedure.
Yours sincerely
Tim Perry – Director of Registry and Academic Services / Secretary to Council – UCL
Gower Street – London WC1E 6BT
telephone – 020 7679 2047 – internal extension – 32047
e-mail – [email protected]
———- Forwarded message ———-
From: clare francis
Date: Sat, Nov 17, 2012 at 11:24 AM
Subject: UCL: suspicions of scientific fraud by prof G Cossu
To: “Perry, Tim”
Cc: James Parry , VC , cope_opsmanager , [email protected], [email protected], [email protected], [email protected], Parker Rosa , Rawle Frances , fgodlee , [email protected]
UCL: suspicions of scientific fraud by prof G Cossu in his own publications
Dear Tim Perry,
Re: “As I understand it, the research in question was conducted before Professor Giulio Cossu commenced employment with UCL (in January 2012).” may be at variance with an official UCL website:
https://iris.ucl.ac.uk/research/personal?upi=GCOSS26
where it states “Appointment Chair in Human Stem Cell Biology Joined UCL 05-Mai-11”
Are both statements correct? Did prof G Cossu join UCL 05-Mai-11? I take “Mai” to mean May.
Would “employment” or “joining” UCL make a difference? Why did you keep to emplyoment?
“You will see too that the procedure expects an allegation of research misconduct to be made by a named initiator (paragraph 1.5)”
Please read: http://publicationethics.org/files/u7140/Flowchart%20Fabricated%20B%20revised.pdf
There is no part about “named initiator”.
Your procedures have got a problem. They go against anonymity (may lead to victimisation), impose more hurdles to setting the scientific record straight.
I believe that you should consult with James Parry at UKRIO.
Instead see this “Thank reader and state your plans to investigate”.
Next “Consider getting a 2nd opinion from another reviewer”
“The initiator shall be required to produce a detailed statement in writing in support of the allegation and submit any supporting evidence that is available to them.” How much more detailed can you get than analysis of the figures in the papers? Did you bother to read my e-mail of Mon, Nov 5, 2012 at 12:20 PM
(please see you own reply of Fri, Nov 16, 2012 at 2:42 PM that includes my e-mail of Mon, Nov 5, 2012 at 12:20 PM which you did manage to send back to me).
I have added my detaled criticisms of the data in prof G Cossu’s papers again because you don’t seem to have read them the first time.
Just look to see if the claims have any basis. That is what matters.
Please see the message directly below to the General Medical Council where there is evidence that prof G Cossu is involved in clinical trials based on his questionable research findings.
BEGININNG OF MESSAGE TO GENERAL MEDICAL COUNCIL
———- Forwarded message ———-
From: clare francis
Date: Mon, Nov 12, 2012 at 11:12 AM
Subject: Re: 1. complaint about behaviour of professor G Cossu head of regenerative medicine at UCL 2. in additional to likely scientific fraud by professor G Cossu
To: “Freeha Akhtar (0161 923 6792)”
Cc: [email protected], [email protected], [email protected], fgodlee
Dear Freeha Akhtar,
Professor G Cossu is the head of regenerative medicine at UCL.
https://iris.ucl.ac.uk/research/personal?upi=GCOSS26
I think he is a danger to the public with a long history of faking his results. It is in the papers he published.
He reported on clinial trials this weekend in London.
Please see:
http://www.actionduchenne.org/r-nav/655.jsp
Action Duchenne 10th International Conference 2012
9th-10th november 2012.
confirmed speakers
– Professor Giulio Cossu UCL London – A phase I/II trial of cell therapy for DMD
Clare
END OF MESSAGE TO GENERAL MEDICAL COUNCIL
Dear All, I will insist here that these queries on published papers would be invaluable to be posted directly on the respective papers at http://pubpeer.com/, please have a look.
It is an excellent platform, and authors and interested readers will be notified of the anonymous peer comments. Anyone can post, anyone can read!
I totally agree with CR. It will be easier to find these comments in the future if they are in a database meant for such a purpose. Hidden away here on a blog message board is asking for them to disappear into the noise.
Un.Believe.Able.
“Named initiator”? “Required to specify”…”what type(s) of misconduct”? “Detailed statement in writing”?
UCL seems to be setting the bar for reporting misconduct so high that no one will bother to reach for it.
I think that showing them a picture, with red lines and arrows, ought to be more than enough. Type of misconduct would have to be given as “unknown” because who knows why the image manipulation occurred until an investigation is made? As for giving your name, that’s going too far. I don’t think the police demand to know your name before they consent to investigate a report of “shots fired.”
“The initiator shall be required to produce a detailed statement in writing in support of the allegation and submit any supporting evidence that is available to them.”
It is enormously revealing statement. Why should the onus be on the initiator (in itself an interesting choice of language – the initiator of a dispute… the trouble-maker, the culprit, the complainer, the aggressor) to do anything?
Surely the onus and the obligation is on the generously taxpayer remunerated members of the research ethics committee to act immediately they become aware someone is abusing the trust of the scientific community?
Its past time that we stopped playing this game. People who report misconduct should not have to play advocate, to push and to plead for institutions and journals to act. They should be tersely alerting institutions and journals to the facts. And then documenting their stone walling and bad faith
I wonder what will happen to those papers published since 2008 which are submitted to the REF assessment? What if they are highly rated in REF, but, after the current round of corrections, later retracted? Same is true to grants, which must have relied in some way on papers published in previous employments.
Of course UCl have another pressure to deal with – UK libel law. I would imagine that it would be quite dangerous for an institution to state publicly that one of its employees was sacked for misconduct. Against a determined individual, this could lead to long and expensive court battle, with the cards stacked very much against the institution. This was my interpretation for the very quiet way Melendez left the University of Liverpool.
I would also note that the current “upgrade” to UK institutional (Universities and Research Councils) policies on research misconduct are simply the consequence of Parliament’s Science and Technology Select Committee asking why there have been so few cases of misconduct in the UK and not one instance of public money having to be returned. It will doubtless come, though don’t hold your breath – medically inadvisable, I believe.
In reply to conradseitz May 14, 2013 at 3:13 pm and littlegreyrabbit May 15, 2013 at 5:50 am
Now you start to see what it like.
Presently I have concerns about many articles by 3 authors at Trinity College, Dublin, but the universtiy does not even bother to reply.
The issue is systemic within the entire global system for “rating” science. I suspect things will even out a bit soon-ish. I am advocating independent and objective evaluations of fields from people NOT within the field and who thus have no CIOs funding-wise….but do posses the expertise required to judge the correctness of the science.
Trinity? Really? Sigh…..(my partner did a PhD in theoretical physics there).
In reply to Dr. Allison L. Stelling May 15, 2013 at 10:02 pm
The first posting about some of the scientists at Trinity College, Dublin, but also elsewhere, has just appeared on Pubpeer.
http://pubpeer.com/publications/CE88BF78111665434427069CBF8CD0
Pubpeer does have the useful feature “Who else should see this?”, which is on the right side of the webpages as they appear.
Plenty more where that came from.
In reply to Dr. Allison L. Stelling May 15, 2013 at 10:02 pm
“I suspect things will even out a bit soon-ish.” That is an interesting thought.
At first it struck me as somewhat optimistic, but I see that you advocate “independent and objective evaluations of fields from people NOT within the field”.
Independent people cultivating the garden of science is really quite a good idea.
The issue I can see is how is that going to happen?
Claire,
from my experience, playing the “deaf & blind monkey” (i.e. ignoring the person who points out misconduct) is the 1st stage of how institutions handle these issues.
2nd stage (if you persist) is playing the “surprised nanny” (i.e. the institution tells you that they are surprised to hear such things and that they take all allegations very seriously, and they’ll ask you to provide as much evidence as you can, which in reality means “show us all your cards, so that we know how to trick you”).
3rd stage is a letter telling you that after an “in-dept, detailed investigation” the investigating body found no evidence for any misconduct (even if the evidence provided by you contradicts to the institution’s own Framework to deal with misconduct).
4th stage is that, should you persist, you’ll be threatened to be sued.
5th stage is that they’ll try to ruin your carrier and life using all possible means.
6th stage is that they might hire a hitman.
I hope that the above will not stop you.
Let me know if there is any way I can do to assist you.
There is a proverb in Europe, which can be translated as:
“The bad guys succeesd only because the good guys do nothing”
Let’s hope that there are some good guys out there willing to do something.
YouKnowBestOfAll,
I couldn’t agree more. First hand, I have experienced all of this except that at stage 4 I have actually been sued (by the fraudster). Thankfully, I have not yet arrived at stage 6. These are the reasons why we need websites such as this one, Science-Fraud, and maybe PubPeer.
This saying is most often written in English along the lines “All that is necessary for evil to triumph is for good men to do nothing”.
Apparently it used to be attributed to Edmund Burke but the closest he wrote was:
“When bad men combine, the good must associate; else they will fall one by one, an unpitied sacrifice in a contemptible struggle.”
Or so it says on wikiquote here http://en.wikiquote.org/wiki/Edmund_Burke
2019 corection.
PLoS One. 2008 Sep 16;3(9):e3223. doi: 10.1371/journal.pone.0003223.
Magic-factor 1, a partial agonist of Met, induces muscle hypertrophy by protecting myogenic progenitors from apoptosis.
Cassano M1, Biressi S, Finan A, Benedetti L, Omes C, Boratto R, Martin F, Allegretti M, Broccoli V, Cusella De Angelis G, Comoglio PM, Basilico C, Torrente Y, Michieli P, Cossu G, Sampaolesi M.
Author information
1
Translational Cardiomyology, Stem Cell Institute Leuven, KULeuven, Leuven, Belgium.
Pubpeer: https://pubpeer.com/publications/3ED214C9172D59B0CF554CE49038F4
2019 correction.
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0220357
After publication of this article [1], the following concerns were raised about the figures:
Figure 2E follistatin and GAPDH panels contain background irregularities such that lane 1 appears to have a different background to lanes 2–6;
Figure 3B Magic-F1 and HGF panels contain background irregularities such that lanes 1–3 appear to have a different background to lanes 4–6;
Fig 5A WT 3d and WT 7d panels appear similar.
Regarding Fig 5A, the authors clarify that the bottom images at 3d and 7d for WT have been erroneously swapped. In the published paper the image below 3d should be below 7d, and vice versa. This is a double IF analysis to test embryonic myosin (red) and laminin (green). Thus all WT upper panels are same images of below using a different light channel.
The authors have provided the original images used in Figures 2E and 3B. The original images do not contain background irregularities and the source of the image artifact is unclear.
With this correction, the authors provide an updated Fig 5, along with the original images underlying Figs 5A, 2E, and 3B (in Supporting Information S1 File).
Reference
1. Cassano M, Biressi S, Finan A, Benedetti L, Omes C, Boratto R, et al. (2008) Magic-Factor 1, a Partial Agonist of Met, Induces Muscle Hypertrophy by Protecting Myogenic Progenitors from Apoptosis. PLoS ONE 3(9): e3223. https://doi.org/10.1371/journal.pone.0003223 pmid:18795097