In November, we reported on two retractions in Cell and the Journal of Molecular Biology involving misconduct in the lab of biochemist Carsten Carlberg, of the University of Eastern Finland, in Kuopio.
Carlberg also holds an appointment in computational biology at the University of Luxembourg, which last year launched an investigation — at his behest, he says — into the case. Well, the officials in Luxembourg finally have spoken, and Carlberg, it seems, may soon be out of his job there, the Tageblatt newspaper reports.
Carlberg, to our knowledge, has never been accused of ethical violations in the case, and the inquiry didn’t change that. Rather, the bad actor appears to be his former post-doc graduate student in Finland, Tatjana Degenhardt. But the University of Luxembourg may seek to dismiss Carlberg after concluding that, as the senior author of the two retracted articles, he bore responsibility for the validity of the results and, by implication, the deception.
Ironically, Carlberg says he asked the school to undertake the investigation even though none of Degenhardt’s research had been conducted using that school’s resources. In an email today to Retraction Watch, he writes:
Although the University of Luxembourg was not directly involved in none of the two retracted papers by Tatjana Degenhardt (neither by funding nor did Tatjana ever worked in Luxembourg), on my request they finally asked a committee to investigate the question of my involvement (since I am employed there). The University of Luxembourg is very young (8 years) and has not yet established procedures for such cases (but will have soon), therefore, it took far longer than probably for other universities.
Carlberg followed that message with this:
taking my responsibility at the moment of the retraction, which the external scientific committee now confirmed, may mean that in worst case I loose my affiliation in Luxembourg. As said before, the University of Luxembourg has not yet any experience with such cases, so they act as formal as they can, i.e. after the scientific committee they will ask now an internal committee (the same that has selected me for the position same years ago), whether my failure as supervisor in spotting the misconduct before publication is severe enough to end my contract. I was asked, whether I prefer to resign before, but I as I am not guilty for fraud, I see this as the wrong sign. Scientifically I am “cool” about the case, but unfortunately somebody gave the internal info to the yello[w] press, who does not understand the details of the case. Since not much happens in a small country as Luxembourg, the newspapers have spotted on the case. Apparently there is some local politics behind it, since the university reported yesterday its annual report (i.e. its success story), which some opponents of the university seem to have chosen to sabotage.
When we first reported on this case, Carlberg had posted on his website a letter accepting responsibility for Degenhardt’s actions. Some readers of this site found that notice refreshingly candid, and even righteous. It has since disappeared from the university’s website, but that seems to be because of lost pages following a merger between Kuopio and the University of Eastern Finland. Today, Carlberg reiterated the sentiment in the letter:
To be honest, I will feel bad about that it happened in my lab until the end of my life, so yes, I feel responsible. Many colleagues told me that it may have happened to them in the same way, since the type of major manipulations that Tatjana did are difficult to spot (minor punctual changes of long Excel tables from PCR runs). As a PI you do not have the time to check all this, you would need 48 h every day with a 10 people lab with 5 qPCR machines running all day. So, I hope that the case of Tatjana is a warning for all other members of my team and others that know the details of the story. She was very talented but finally too ambitious. It gave me a big learning experience and I reduced the number of team members as a consequence.
Correction, 5 p.m. Eastern, 5/21/11: “Post-doc” changed to “grad student” in headline and third paragraph. Thanks to commenter thetati for pointing out the error.
Please see an update with a few more details.
Hat tips: Corto for Tageblatt story, Nancy Lapid for translation
I disagree about Carlberg’s statement regarding the “major manipulations”. Many of them did not invole qPCR data, but images of bandshift assays and a 3C assay. Whole gels (Degenhardt et al., JMB 2009) or single bands (Degenhardt et al., Cell 2009) were re-used several times, and for the bandshifts, different quantification data were reported for identical gels when compared to previous publications (Degenhardt et al., JMB 2007).
My contribution to this debate would be to urge some caution and instill some balance. I agree this episode is deeply disturbing and requires a full and open investigation that comes to a final conclusion. What is not helpful is that Prof. Carlberg continues to be prosecuted by innuendo in the court of public opinion.
It is worth highlighting Prof. Carlberg’s achievements in science over a period of more than two decades. This has resulted in a sustained and high quality level of publication, with over 150 primary research papers, reviews and book chapters. This output has had, and continues to have, a major impact on the field of nuclear receptor biology. As a result, his work is extensively quoted, with more than 10 papers cited over 100 times. Importantly, he has attained an h-index of 40, indicating a very broad citation base by his peers. This is a critical fact. If any of the papers were fraudulent they would not continue to be cited by his peers. This fact is probably the greatest honor a scientist can receive and should not be dismissed lightly. Each time a paper is cited it forms part of ‘logically articulated structure of ideas which is already, though not yet half built, the most glorious accomplishment of mankind.’ (P. Medawar, Nobel Laureate). In each citation another scientist has independently established links between themselves and the cited work. This is the ultimate of validations.
Against this peer-reviewed validation appears the inability to draw a line under this episode and instead continuing to flame it. This reflects poorly on the host institute and wider community. I would urge everyone to display some balance and caution in these comments.
I quote here the “mission statement” of this blog: “Tracking retractions as a window into the scientific process”. This encompasses not only ethics and personal responsibility but also peer review and institutional processes. There is no doubt that Carsten Carlberg has made significant contributions to the body of scientific knowledge. In fact, it can be argued that he is one of the star faculty at the University of Luxembourg, a small and new institution that is still trying to make its mark on the international scene. It is clear from such statements as have been made public that uni.lu and the faculty of the Life Sciences research Unit are now dropping him like a hot potato, presumably concerned that his sullied reputation will rub off on them. This against a background where it appears that contrary to other recent cases (see e.g. the Bulfone-Paus affair) the accused has been making an effort not only to clear his name but also to address the issues head on. I certainly hope that as this story unfolds the full extent of what happened and why becomes clear. Were the (so far anonymous) researchers supporting him misguided, or did they have factual reasons for their action? Did the investigating committee really uncover new evidence that has not yet been made public? Is this a case of infighting in a small academic department?
This being said, I do not believe that a stellar academic record provides evidence for a lack of wrongdoing. That would really be naive.
I am wondering what is exactly the scientific contribution? What do 150 primary research paper mean when nobody knows how those data for the 150 paper were created? How can Pondera know that these are high-quality papers? And just the fact that those paper were citied doesn’t tell anything about the quality of the paper. It may be also important to see the context when those primary publications were citied. – All the comments especially in the luxembourgish press, are questioning the general way of Carlberg to perform research. It is not a discussion about a PhD-student who manipulated results in 2 publications, it is the question what is the quality of Carlbergs research? That means that one need to question each of those 150 primary research papers.
Instead of judging researchers according to the number of primary research papers, we should either start to be doubtfully with persons who have an astonishing high number of publications.
The case it is not really all about PhD student (she was not post-doc)
see
http://www.journal.lu/2011/05/20/%E2%80%9Eder-fall-ist-komplex%E2%80%9C/
As a parallel to the Bulfone-Paus case, 16 other professors supported Carlberg in a letter to the university as reported in the article posted above. I find this reaction quite surprising since the manuscripts were published despite grave violations of scientific principles. Fame and other personal impressions unfortunately cannot predict scientific quality. Scientists should be very well aware of the fact that opinions can easily cloud one’s perception and stand in the way of an objective conclusion.
Your title seems a little bit misleading. If one can believe the local press paraphrasing the head of the university, Carsten Carlberg risks to lose his post not because of “Grad student’s misdeeds” but because there were serious problems inside his lab.
“It gave me a big learning experience and I reduced the number of team members as a consequence”
Sounds good, but to my knowledge it was not Carsten Carlberg who reduced the size of his group, but the group members by themselves requested to leave the Carlberg lab for scientific reasons.
I agree with benoitmajerus, the title is misleading, since the investigation in Luxembourg was obviously not only about the Cell paper, but rather initiated by this story.
Indeed, the title is misleading. The summary of the international commission’s report was broadcasted in Luxembourg making it likely that the retracted papers are just the tip of the iceberg.
It is a sad commentary for science and society in general that we look for scapegoats other than the perpetrator. Usually it is the PI who fakes the data and the grad student gets the blame. Here it is the reverse, Tatjana Degenhardt seems to be living happily ever after at MIT in the lab of Ernest Frankel.
http://www.hunter.cuny.edu/qubi/repository/files/QUBI-MIT_.pdf
It looks like she is not listed on his website or in publications.
http://fraenkel.mit.edu/people.html
She’s mentioned on his CV, with 2009-2010 as the years she was there.
Thanks for the comment, Azzy. We noted Degenhardt’s time at MIT in one of our earlier posts on this:
http://www.retractionwatch.com/2010/11/12/cell-retraction-for-bogus-images-in-genetics-paper-reveals-another-in-journal-of-molecular-biology/
A procedure to fire Carsten Carlberg was actually initiated, according to a press release from the University of Luxembourg dated May 20th. The press release (in French) is here:
http://wwwen.uni.lu/university/news/press_releases/20_mai_2011_apres_le_rapport_d_une_commission_d_enquete_procedure_de_revocation_d_un_professeur_en_bioinformatique_entamee
Here is a rough translation:
“In accordance with the arrangements stipulated by the Law on the University of Luxembourg, a procedure for dismissal has been initiated by the University against one of its professors of bioinformatics.
The professor in question had to retract two publications because data had been falsified. One of his co-authors, a graduate student at another university, was responsible for the falsified data. On the basis of this information, the University decided to verify the scientific work of the professor.
The international expert committee mandated by the University has concluded in its report that while no intentional manipulation of data could be blamed on this scientist, he nevertheless bears responsibility for the retracted papers. Additionally, the committee concluded that he did not mentor or direct his research group and his doctoral students in Luxembourg according to internationally recognized standards. The committee has also expressed reservations regarding other papers that he published.
During its meeting of 16 May 2011, the Governing Council of the University decided to initiate the dismissal procedure of the professor in question, as stipulated by law. As Rector Rolf Tarrach underlined, this is an isolated case. “it is our duty to protect the scientific reputation of the young University of Luxembourg and that of its researchers. Beyond international standards, the University conducts research according to its own code of ethics and has an obligation to take adequate measures when its rules have been violated.”
So it does look like the two retracted papers are not the full story, assuming of course that the University’s representation of the conclusions drawn by the expert committee is accurate. Knowing a little about the U. of Luxembourg I believe that they are extremely concerned about their reputation and credibility, and would not make such a statement unless it was well founded.
#carlberg’s department at #unilu distances itself from #carlberg http://goo.gl/tCAfL
Thanks Ivan, I read your other article later.
I am also concerned about the pack mentality shown repeatedly in science. Although it is refreshing to see them called out. It reminds me of the Baltimore/Imanishi-Kari trial where so many scientists defended obvious data fakery to the point of hysterically attacking Dingel, the congressman leading the investigation. In that case the whistle blower was lucky to find any job at all and apparently left science, while Imanisi-Kari is still working at Tufts.
The 16 colleagues who have declared their support for the professor in bioinformatics in a letter to the university are external people not affiliated to our research unit. We are committed to highest scientific standards in our daily work and would like to distance ourselves from the incorrect working practices described in connection with the recent affair concerning this professor
Azzy, the second paragraph in your comment was copy/pasted from the the Life Sciences Research Unit’s page referenced by Benoît. Does this mean that you endorse this? It looks to me like solidarity with colleagues is not the order of the day at uni.lu. This whole affair gives me a funny feeling, given that Carlberg himself seems to have asked for an inquiry to begin with.
Where does it say that Carlberg asked for an inquiry? My interpretation of the articles is that it was started because of “irregularities”, these being the retractions and other things. This article http://www.lessentiel.lu/de/news/luxemburg/story/27324372 says that a grant application for the Fonds National de la Recherche (FNR) was somehow connected to “incorrect methods”.
Carlberg told us that he requested the inquiry. From our post above:
Honestly, if a retraction is coming to public knowledge, would it not it be a good tactical move to “ask for an inquiry”, as most likely it would come anyway?
Note: This comment’s third paragraph is a quote of comments on the U of Luxembourg’s Life Sciences Research Unit (bottom of page): http://wwwen.uni.lu/research/fstc/life_sciences_research_unit
Just wonder what is the connection of these 16 colleages with Carsten Carlberg, are they possibly friends, collaborators, co-authors or close in some other connections?
I don’t have any association with this case or any other cases that have appeared on retraction-watch, but an academic myself, I think the maintainers of this site should be ashamed of themselves for making money based on malicious gossip within the scientific community. Truly a parasitic enterprise and yellow journalism at its worst. Anyone who is listed on this site should immediate consider legal action. I challenge the maintainers of this site not to delete this post.
Thanks for your comment. We always appreciate feedback. Generally, that feedback has been quite positive: http://www.retractionwatch.com/what-people-are-saying-about-retraction-watch/
As far as “making money,” while we’d be happy to generate significant income from the site, our total revenue to date is in the mid-five figures — but only if you include the cents. That’s from a column we write for LabTimes. We blog at Retraction Watch because we believe in transparency.
In the meantime, if you have particular facts we’ve published that you know are inaccurate, we would welcome the opportunity to correct the record, so please get in touch.
Yeah, I’m an academic and I demand not to be held accountable for my actions!
As an (ex-)academic, I’d say the parasites and cowards we need to watch out for are the members of that community who hide their misconduct behind the (often deliberately) ineffectual infrastructure that is supposed to investigate it. Retraction Watch is a welcome public service that brings some much-needed sunshine into the darker recesses of scholarly life.
Where’s the Like button when you need it?
Could agree more with you, Bill.
I couldn’t disagree with you more, gr. I am astonished at the way some journals and scientists feel that fraud should be treated. As somebody working in academica, I applaud the intiative of Ivan and Adam (who I suspect are not getting rich by making this site) because Retraction Watch brings some much needed transparency in the -often grubby- process of retractions.
Sure we need transparency but we also need to keep a balance. Yes scientific misconduct is the antithesis of scientific inquiry. A website dedicated to reporting on retractions can develop a world-weariness in which everyone is presumed to be equally guilty. In this case it does not appear to be justified.
Or it can promote hope in those who do not cheat that perhaps the cheaters will be exposed one day. What about the world-weariness and cynicism of those who see cheaters and fakers profiting handsomely by their amazing “successes”?
I don’t have the impression that this website is claiming Prof. Carlsberg is guilty of misconduct. So far Adam and Ivan are reporting on the somewhat opaque statements by the Univ. of Luxembourg and the popular press in Luxembourg.
If he would be fired from his position in Luxembourg only for fraud by a post-doc in his lab in another university, then it would be a case of overreacting in my opinion. But maybe there is more? Enough to report on I guess.
Maybe the reporting becomes the story? It seems that a process to dismiss Prof. Carlberg was started (May 20th. May 16th)? Why has that started? If there is a report that shows unequivocally that he is guilty of misconduct then his host-institute should release that document, be proud that they are acting decisively and dismiss him. If they do not have this evidence then they should support their employee and work out what is a sensible way for all within the unit to go forward and be proud of their measured handling of a difficult issue. At present they appear guilty of doing neither and instead allowing the whole affair to drag on in an over-dramatic manner. The motives of the host institute in this regard appear dubious.
Given the large diversity of lab types just within single -ology disciplines within the US…I would REALLY like to know more about alleged international standards for running a lab.
Ivan Oransky and Adam Marcus are doing the wider public a general service. I quite like it when they point out the inconsistencies of the authorities, which are management primarily. They are allowed to read the official texts put out by managements and comment on them.
I suspect that many people shy away from reporting scientific misconduct because it is seen as “rocking the boat”, i.e. will put governments and investors off science (less money will be forthcoming), when in fact scientific misconduct is a fact of life.
I imagine that the prospects of promotion are dimmed if you get a reputation of “rocking the boat”
Ivan and Adam are in the medical and academic worlds so they do know what to look for.
If I wore a hat I would take it off to them.
Keep up the good work!!!
There is a piece of official information about the case that has not yet made it into this debate. A press release dated May 19th from the National Science Fund reads as follows:
“Luxembourg, le 19 mai 2011. Dans le cadre de l’évaluation d’une proposition d’un projet de recherche, le FNR (Fonds National de la Recherche) a été averti par un de ses experts externes qu’un chercheur travaillant au Luxembourg est co-auteur d’articles pour lesquels des problèmes concernant l’intégrité de la recherche ont été signalés.
Dans le cadre de sa mission de développement de la recherche dans le secteur public en termes de qualité, le FNR a soulevé des interrogations dans certains projets du chercheur en question et il a consulté des experts en la matière pour vérifier ces points afin de lever les interrogations.
Face aux problèmes constatés par ces derniers, le FNR a décidé de ne plus mettre des fonds publics à la disposition des projets du chercheur en question tant que les doutes et interrogations invoqués n’auront pas été levés à la complète satisfaction du FNR.”
I can provide a translation if needed, but it looks like the FNR has conducted its own inquiry and also found problems that go beyond the publication of the two papers with falsified data. Whether this inquiry involved experts others that those mandated by the University is not clear. The FNR has suspended all funding for the lab until the problems have been resolved to its own satisfaction.
Prof. Carlberg’s publication record can be seen clearly and openly with the number of citations and the timing of those citations publicly available. If an article is cited by other workers in the field then the most logical explanation is that they see how that work or perspective fits with their own studies. In the current case it is unclear how that could be manipulated. An H factor of over 40 means he has 40 papers that are cited at least 40 times. That is a very significant contribution and is clearly in the public domain
I still do not see that there is published evidence concerning any doubts about the quality of any other papers other than the two that were retracted. Again, if the host institute or FNR have definitive proof they should make it openly available to the public and the readers of this website. Otherwise this discussion appears based on a little fact (the two papers being withdrawn) and a lot of innuendo.
Well, pondera, If you actually do research in field which is somehow competitive — e.g., nuclear receptors, MYC, P53, TGFbeta, whatever — you will quickly notice that there are people who will continously put out many papers and/or many papers in high-profile journals. It has been my personal experience in several fields that some labs’ results are reliably reproducible for me, whereas others are not. For some labs’ results I can only comment that none of them were ever reproducible for me or my colleagues. These observations do not correlate at all with the citations those papers get. You could hypothesize that this is our inability to perform some types of assays, but there is more evidence from other labs we know closely that this is not the case, at least not for the majority of assays. I might add that one comment by a reader on the Carlberg affair on this site was that Carlberg is “a name in the field”, whatever this may mean. To cut a long story short: In many fields, it is well known whose results you can trust. I know a lab which publishes so well that when you see it from a perspective of impact points, the PI is among the top ten researchers in the US. Nevertheless, most people say that they can almost never reproduce their results. Nevertheless, those papers are cited often. There is no rule that you can only cite a manuscript if your lab reproduced its results. I think your hypothesis about citations and scientific quality cannot hold true; many if not most scientists cite manuscripts they have not tried to reproduce, and in some cases, not even read thoroughly. Just an example: In the last years, There were three papers in Cell directly connected to our research, and they also got quite a number of citations. For one, we could not reproduce some key findings. Two were nonreproducible in toto. One has been retracted meanwhile. Q.e.d.
Although your first line is an argument ad hominem I will indulge you. I’m a research scientist and group leader working in a biomedical field and in an academic setting. I have about 20 years experience. In all honesty from my experience, and I’m not trying to be personal, if you look for demons you will see them. So generally I think workers do the best they can to do the most insightful and reliable work. I think genuine deception is still very rare considering the sheer volume of papers published.
As i consistently have said here (and I’m not sure what more i have to say!) if the Luxembourg or FNR have definitive evidence of a wider deception then they should release it, otherwise we are merely re-cycling gossip and innuendo.
To the point of citation and quality. Broadly it’s true; that is, peer review and publication is the least worst system but in my experience, work that is good becomes recognized and cited. Of course there are exceptions. I think the most interesting paper i did as a post doc, published in 1996, has still only been cited 10 times and clearly was ahead of its time 😉
I sincerely apologize for coming across too fiercely, Pondera. My personal experience may not be representative, but it has been largely negative when it comes to high-profile journals and the groups which frequently publish in those. Of course, there are many honorable people in the labs out there. But I think fraud is much more common than scientists and the public would like to think. A former colleague of mine was hired by a big pharmaceutical company; when he started, they told him to always reproduce key experiments from the literature if he was starting projects because of them. They also told him that the company’s average success rate in reproducing results from publications was 30%. Of course, the remaining 70% are the sum of technical reasons, so-called cherry picking and outright fabrication of data. No hard feelings, please — but I still estimate that a two-digit percentage of published manuscripts in life science contains bogus data.
no apology needed!
Carlberg has active research group at the University of Eastern Finland. See:
http://www.uef.fi/biotieteiden-laitos/tumareseptoritutkimus
You will get the page in English by clicking “EN” above the title “tumareseptoritutkimus”
He is indeed paid also by the University of Eastern Finland (50% of full time salary)
Quoting Carlsberg: “As a PI you do not have the time to check all this, you would need 48 h every day with a 10 people lab with 5 qPCR machines running all day.” If this is so, then it is high time that computer security provides laboratories with the means of tracing changes made to experimental data. I’m not exactly in lab soft but I’d guess such systems exist. Would it not be the PI’s responsibility to implement such systems beforehand?
Or maybe the fact that Carlsberg is working full time (100%) for University of Luxemburg and half time (50%) for University of Eastern Finland is certainly very interesting end of each month, but leave less time to perform the right controls to assume a correct work ?
Totally agreed.
Public availability of primary data would be a good start and would likely deter most grad students from fabrication of data. Scans of complete blots instead of the tiny cut-outs usually published. Lists and the curves of the raw qPCR data, or just the raw data file, instead of the nice diagrams with variation indicators. Etc. In times of nearly unlimited online space, what is the problem with adding all this to the online supplemental material?
Fabrication of the raw data, expecially of large data sets, is somewhat difficult, if you want to have statistically reasonable results (not too good to be true….).
Regarding all data coming out of machines, ultimate control for the PI would indeed be to have primary data automatically saved and the system only giving copies out for work. In fact, I know of labs were this is the way it works for some of the machines.
If the entire lab or the PI fabricates, all this will not help with 100% certainty. But it would help deter a lot of fraud.