MIT lab retracts Cell synapse tagging paper for falsification or fabrication

cell feb 2013A rising star at MIT has retracted a paper after an investigation found that her former postdoc had “falsified or fabricated figures.”

Alice Ting, winner of an NIH Directors Pioneer Award and named one of Technology Review’sInnovators Under 35,” published the paper, “Imaging Activity-Dependent Regulation of Neurexin-Neuroligin Interactions Using trans-Synaptic Enzymatic Biotinylation,” in Cell in 2010 along with Amar Thyagarajan.

The notice is refreshingly detailed given the circumstances:

This article has been retracted: please see Elsevier Policy on Article Withdrawal (http://www.elsevier.com/locate/withdrawalpolicy).

This article has been retracted at the request of the Author.

This paper introduced a new methodology, BLINC, for detecting the trans-synaptic binding of neurexin and neuroligin proteins, and applied BLINC to study the interaction dynamics of these proteins in neurons. Since this publication, my laboratory has found that BLINC cannot be reproduced in neurons using the constructs and protocols described in this paper. After I brought forward concerns, the Massachusetts Institute of Technology conducted an independent investigation. Communicating the findings of that investigation in a letter to Cell, Dr. Claude Canizares, Vice President for Research and Associate Provost, stated that: “MIT found that the first author, Dr. Thyagarajan, falsified or fabricated figures in this publication. MIT’s investigation also found that Dr. Thyagarajan was solely responsible for the scientific misconduct that resulted in the falsified or fabricated data.” I therefore wish to retract this publication. My laboratory has subsequently found that, with modified constructs and protocols, BLINC can be used to detect trans neurexin-neuroligin interactions in neurons. We will report this in a future publication. I deeply apologize to the scientific community for any loss of time or resources caused by this publication.

The first author, Amar Thyagarajan, has declined to sign this retraction notice.

The paper has been cited 17 times, according to Thomson Scientific’s Web of Knowledge. When it was originally published, Autism Speaks, which funded Thyagarajan’s postdoc, heralded it:

The ability to observe the active development of synapses will undoubtedly factor into future discoveries, paying dividends for some time to come.

MIT called it “a better way to see molecules at work in living brain cells.”

Thyagarajan is listed as a technology specialist at Clark + Elbing, a patent law firm in Boston. We’ve tried reaching him for comment, and will update with anything we learn.

Update, 1:30 p.m. Eastern, 2/15/13: Thyagarajan sent the following comment:

I was not contacted by Cell about their decision to retract the paper.  I want to be clear that the retraction was done over my objection.

I stand by the data that was published and the methodology that I developed.  I and others have reproduced this method over four years.

The findings against me were the result of a deeply flawed and sloppy investigation that ignored evidence that someone had tampered with and deleted my data, after the publication of the paper, and made it look as if I had falsified data.

I am greatly disappointed that Cell issued the retraction without contacting me and giving me an opportunity to explain the facts and my side of the story.

The matter is now at ORI. I expect to have a full and fair opportunity to be heard before impartial fact-finders and am confident that my innocence will be established.

Update, 8:30 p.m. Eastern, 2/16/13: Thyagarajan tells us he has resigned his position at Clark + Ebling.

Hat tip: Christophe Leterrier

90 thoughts on “MIT lab retracts Cell synapse tagging paper for falsification or fabrication”

  1. Here goes one more “postdoc” was solely responsible for all the data. Come on world, get real. Is it possible that one person is responsible for all the mess??? Come on Alice Ting, take some blame.

    1. Wait, she requested for the retraction, quite noble and unusual. That are certain things that it is impossible for the PI to discover. I’m not saying that she is 100% innocent but the request for the retraction indicates that way!

      1. A good PI should always be on top of her game. Looks like this is a difference in opinion between a PI and the post doc. Why else would she publish a similar paper, at the same time as the retraction and yet apply for the patent almost 2 years prior. Something is fishy!

        1. There is a patent application on the lipoic acid ligase technique from 2007. Then there is an application extending the technique to this new purpose, filed in 2011. And, under international patent law, you always have to file the application before publishing the result or else you lose the ability to get a patent.

          What is probably the real concern is that — if they really believed the BLINC technique was faulty — they held the retraction until the replacement study was published. How many people tried and failed to reproduce the technique between the time they discovered the problems and the time they published the response.

          1. Hypothetically, the 1st author may have seen the patent when it was published in 2012 and felt the ID-PRIME technique it described as one of the applications contained intellectual input from him, since he had already set up the system using biotinylation. He contacted the last author accordingly.

            The last author then has an incentive to undermine his work.

            Nothing would surprise me less than a University would deny due process to a junior ex-member of a lab – whether whistleblower or accused. The first author has given his scenario what he believes happened and deserves a hearing from a body without such an obvious conflict of interest and be given full cooperation by MIT to prepare his submission

    2. You are incorrect to imply that the PI of a retracted paper will not suffer any negative ramifications, especially a junior PI. Do you think her new grant applications and manuscripts are going to be reviewed in ignorance of this retraction?

  2. So you were not able to find out that the figures were fabricated but you were able to “subsequently found that, with modified constructs and protocols, BLINC can be used”. You did it yourself this time, or was someone else? How can you tell that the figures were not fabricated a second time? Why are you the corresponding author?

    1. Isn’t this pretty straightforward Anonymous? Dr. Ting’s postdoc fabricated data in a manner that wasn’t readily apparent. Since research work in a lab is cumulative it’s not surprising that subsequent work built on the (now-retracted) study. The students/postdocs doing this found that the experiments couldn’t be reproduced. There was some investigation of this (perhaps it was determined that the constructs/protocols simply couldn’t have produced the reported observations, indicating that chicanery, rather than error, was involved). In the course of the follow up work new constructs/protocols were found that worked.

      It’s a sort of microcosm of the way science “self-corrects” isn’t it? If an observation is important enough (and it always is in relation to the progression of a lab’s research), it’s likely to be reproduced. Of course it’s horrible to find that your lab has produced a major howler. However much better to be the one that discovers the error than to have someone else do so….

      1. I would love to see fabricators and fraudsters get stripped off of their titles and awards, and pay back all the benefits they had received before getting exposed. What about those who publish similar studies under different names (different spellings to deceive editors)in lower tier journals ? How could we expose them especially that they are now holding higher positions in Academia? Is there a way to purify science around the world and clean this mess?

  3. Why should Dr. Ting “take some blame”? Her group discovered that the BLINC method using the constructs and protocols described in the paper can’t be reproduced. She therefore retracted the paper. In what sense is she blameworthy?

    Of course in hindsight one can always regret that specific experiments weren’t independently repeated by someone else in the lab, or that one didn’t look over the shoulder of the postdoc doing the experiments. Trust in the honesty of one’s colleagues is required in order for a lab to function; unfortunately this can turn out not to be justified on occasion. At that point you may be left with the unpleasant tasks of obtaining an independent audit, examining experimental lines to recover what is valid and establish what isn’t, and even retracting paper(s). Dr. Ting seems to have been both unlucky with misplaced trust but also praiseworthy for addressing the problem maturely.

    1. then how does one explain the similarities in the two papers and what ever happened to the filing of the patent. why isnt Dr T’s name on the patent.

      1. Because, even though the goal of the two papers is the same (detect where two cells make contact), the detection technology is different.

        The Ting lab applied for a patent in 2007 for a method involving lipoic acid ligase.

        Dr. Thyagarajan developed a different method involving biotin ligase. No patent was ever applied for this method. (Possibly it was not believed to be a novel enough modification of prior art to be worth applying for.)

        The Ting lab now says that Dr. Thyagarajan’s method doesn’t work as published, that even an improved version doesn’t work very well, and that he may have fabricated the data showing that it did work. Instead, they developed a modification of their 2007 technique that works better, and then filed a new patent application as an extension/amendment to the 2007 application.

        The Ting lab may have a financial incentive to say that the new method is better, but I’m not sure what their motive is to say that the biotin method was fraudulent. Since the two technologies are different and the lipoic acid ligase method was patent-applied for in 2007, there does not seem to be a “prior art” reason to knock down the biotin method.

        The matter seems to be in the hands of the ORI.

  4. Are those rhetorical questions Ressci? Why should there be “records” on the webpage? I don’t have records of all my former postdocs on my lab webpage. You can put whatever you like on your lab web-page. You could answer your question about the sex of the postdoc (“himself/herself”) by reading the top article carefully.

    1. Absolutely you are right. It is PI’s decision to include a name on his/her webpage or not. You win.. I am surprised that this particular postdoc’s name is missing. That’s all.

  5. Surely a PI has regular meetings with postdocs and grad student to look over data, check out signal-to-noise and so on? If not then there should also be an apology for poor lab management.

    1. Look at the data presented in the paper. It is mostly fluorescence microscopy pictures and image intensity values calculated by software on those photos. So what the PI would see in those meetings is graphs, spreadsheets and photos. A photo can be labeled with any labels, and as we know from that social psychologist (too lazy to look it up), spreadsheets can easily be faked. The only way the PI could guard against this is to actually perform the image quantification herself (assuming that the photos were taken correctly) or take the photos and perform the quantification (assuming the photos themselves could be mislabeled).

      If you want to run a lab that way then fine. But understand that a PI who has to verify every experiment personally by hand, will have very little time to teach, write grants, or train students.

      1. Strong dreams wrote: “But understand that a PI who has to verify every experiment personally by hand, will have very little time to teach, write grants, or train students”

        Any good PI will go through data with their staff/students.

        The PI is responsible for the data, legally and ethically.

        1. “Going through the data” can detect most cases of sloppiness and corner-cutting in the preparation of figures (like spliced Western blots or suspect blot quantitation). “Going through the data” may not detect outright fraud depending on the data format. I, as a PI, can certainly ask to see original Western blots to make sure there are no improper manipulations in the construct of a figure. I, as a PI, am probably not going to ask to sit at a microscope and view hundreds of original slides that were used to generate a table of image values. If you can’t trust your students and faculty to be ethical (until proven otherwise), you probably shouldn’t ever try to run more than a 2-person lab.

          I know many faculty at small “teaching colleges” who have labs and do research. For the most part they have no staff, or at most one technician. The real purpose of the lab is to provide internship opportunities for undergrads. The faculty mostly do their own experiments, and they rarely publish (nor are they expected to by their institution).

          That’s a fine way to run a lab if it suits you, and it will certainly ensure that the PI can personally vouch for every piece of data.

          But most labs don’t run that way, and an awful lot of science would never get done that way because many projects are complex and need more hands, brains and bodies at the benches. No matter how good I am, there are some experiments I can’t perform by myself. As soon as you bring in a second person, whether a student, research fellow, or faculty colleague, you have to start trusting them (until proven otherwise).

          There is, sadly, one person in my department who I do not trust, and who I would have to scrutinize quite closely if he/she ever proposed a collaboration. But I took that same approach with everyone, I wouldn’t get any work done.

          1. Strong dreams:

            As a PI you are no doubt aware you are responsible for the work, and accuracy of ALL the work, from your projects. You also reap any financial, IP and prestigious rewards, promotions etc.

            Such rewards can be very considerable.

            PIs rightly get the rewards, and they are not shy in claiming said rewards, and yet you now claim bare few of the responsibilities if fraud occurs.

            I would argue the two are inextricably linked.

            We all know it is not about a PI performing experiments, it is about the PI showing due diligence in data analysis, and teaching their underlings sound scientific practice.

            Without due diligence how does any decent PI show they have spent grant funds appropriately?

        2. “The PI is responsible for the data, legally and ethically.”

          Not really. Of course when I as a PI come to present data from my lab in a paper, or grant application or grant final report then I do take responsibility for my use and presentation of the data. Pretty much like Dr. Ting did when she discovered a serious problem with some of the data she published.

          But students and postdocs also have a responsibility for their data. In fact one of the criteria by which we assess the development of a PhD student is the extent to which they develop a proper responsibility for their own work.

          1. Yes. Dr. Ting could just have easily submitted a new paper on “An Improved BLINC Protocol for Detecting Trans-Synaptic Interactions” that would have fixed the problems but without admitting there was fraud in the first version.

        3. @Stewart, February 15, 2013 at 2:17 pm:
          I think the issue is with the definition of “responsibility” and the appropriate consequences.

          Dr. Thyagarajan actually committed the fraud. Dr. Ting failed to detect it. What are the proper consequences for that? For one thing, she retracted the paper, which is enormously embarrassing even if the actual fraud was by a student or associate. For another thing, her lab spent thousands of hours tracking down the problem, explaining it, and correcting the scientific record.

          Do you want more, and if so, what and why? Certainly there seem to be some labs that have a pervasive problem with fraud, or at least inappropriate data “adjustment.” And for those labs sanctions might be appropriate. And there are some PIs who themselves commit fraud, and who should be disbarred period end of story.

          But I, for one, recognize that even a diligent PI can be victimized by a determined fraudster. Disbarring or firing the PI is only a further victimization and does nothing to improve the conduct of Science.

          1. “Yes, PI is responsible for data legally and ethically.”

            That’s the second time this has been asserted and I have to say I don’t really know what it means. I had a look at the various codes of conduct that I sign up to as a PI in my employment and nowhere does it say or imply that I’m “legally” and/or “ethically” responsible for data.

            There are certainly responsibilities that I sign up to that involve safety and good practice in the laboratory and I certainly have responsibilities in relation to the facilities that support students and postdocs on projects and also in relation to contracts with funders that work is undertaken in line with agreed programs and timelines if applicable….and so on. I have responsibilities in relation to the training and mentoring of students (this may be somewhat devolved if my department has a graduate school in which some more centralised aspects of student training take place).

            But students and postdocs have responsibilities to. They also sign up to codes of practice in which legal and ethical aspects of laboratory work are adhered to. If a student or postdoc acts against legal or ethical codes then it is they that have broken their agreement…not me. If a postdoc fabricates some data, and this is unwittingly included in a paper, I am neither legally nor ethically responsible for that fraud. However if the fraud comes to light then it is certainly my responsibility to address the problem.

            Perhaps it would help if you (Crux) or Stewart would give an example. It sounds to me like you are arguing that if a student (say) fabricates some data and this is discovered, then it isn’t the student that should be dismissed (which is what might well happen in my experience), but the PI since s/he is “responsible for data legally and ethically” (whatever that means). But that can’t be right…so can you give a better example or two?

          1. Agreed Crux.

            A PI who doesn’t look at data coming from their laboratory is failing to accept their responsibility, which is a failure to lead.

          2. hmmm…the nesting on this thread has caught me out and my response to your post has appeared just above yours instead of below….

          3. Imagine the following scenario:
            “You tell your PhD student of Postdoc we need this data in two weeks”
            The PhD student of Postdoc tells you that is impossible, you reply do it.
            and guess what you get the result you so dearly wished for…

            Do you think hierarchical pressures do not influence these events. Worst of all they are very hard to prove and quite honestly safety nets for those in the position of power (PIs) to easily cover them self behind the curtain and say it was not my fault and point the finger at postdocs.

            So then who is responsible you for putting your employees in a impossible situation or they for getting what you wanted.

        4. LEGALLY, the PI would not be held responsible for research misconduct of his/her trainees unless it could be proven that the PI had a direct hand in the falsifications or fabrications at hand.

          1. Chris

            Certainly, any post-doc or student, or techinician or whoever else does science-fraud they are indeed responsible. But they are not alone.

            I would also suggest PIs whom go through data very carefully prior to publication are equally responsible.

            Take a scenario – a student comes in with “amazing results” – the PI sees “stars” consisting of the Nature paper they always dreamed of, and the magic word “a big grant”.

            Now imagine, just for a second, that the PI has suspicions about the data, but lets it go, publishes and gets the grant. All is well and good, right?

            2 years later – suspicions are raised and an investigation shows “errors” – a correction occurs (or even 2 corrections), and perhaps a retraction. Others dig deeper and the PI has several papers, that do not involve that student, under suspicion.

            Is it the student or the PI who bears responsibility?

            Is the PIs failure to show due diligence accidental?

          2. “Certainly, any post-doc or student, or techinician or whoever else does science-fraud they are indeed responsible. But they are not alone.”

            They might well be alone. If the student or post-doc had falsified the data in a manner that was not apparent to anyone else (say at the time of submitting a paper) then the student or post-doc would be entirely responsible.

            “I would also suggest PIs whom go through data very carefully prior to publication are equally responsible.”

            Not necessarily. If the student or post-doc had fabricated data in a manner that wasn’t apparent, the PI might well have gone through the data very carefully without identifiying the fraud. They wouldn’t bear “legal” or “ethical” responsibility for the fraud.

            Perhaps we can agree that those who are responsible for a fraud are the ones that are responsible for the fraud. In circumstances that a student or post-doc perpetrates a fraud alone without this being apparent to others then the student or post-doc is entirely responsible. Of course if the PI subsequently discovers the fraud then they (the PI) is responsible for dealing with the resulting problem. But the responsibility for the fraud lies entirely with the perpetrator.

            In your scenario in which the PI has suspicions about the data, but lets it go then the PI would share responsibility. The PI would be acting unprofessionally and probably unethically. That would be a difficult scenario to pin down investigatively since it would require accessing the mind of the PI if there wasn’t discrete evidence of the PI’s suspicions.

            The point is that there is no blanket legal and ethical responsibility for data on the part of a PI, and the situation depends entirely on the circumstances. If a post-doc or student prepares a fradulent set of data that fools the PI and that ends up on the paper the PI bears neither ethical nor legal responsibility for the fraud. If the PI colludes in the fraud as in your extra scenarios then s/he bears responsibility.

          3. To Chris a 4:27, you are entirely correct (at least in the U.S.). If you were to review all of the federal register notices coming out of ORI you will find that PI’s or supervisors are found culpable only if they were aware of the falsification (or directed it) even if they themselves did not have a direct hand on the data. Being a bad mentor or boss is not misconduct as far as law is concerned.

        1. Is a school principal responsible when one of the teachers starts sexting a student? Is the police chief responsible if one cop is caught planting evidence? Is the President responsible if a cabinet official does something shady? Your argument boils down to no human enterprise should ever involve more than two people who can cross-check each other constantly.

          1. I agree that you should check out every blot direct from the film and I do that for my students. How do you do whether the student treated the cells/animals in a different concentration he/she said he did? How about if he/she put a sample saying it is A and it is B? If you know a way to do that, please let me know how and I’ll be more than happy to do it in my lab. That is a limit for everything. I do have to rely on trust at some point or do everything by yourself.

      2. I guess, if you do such software calculations blinding would be an option, if not even requirement. There are many ways to improve experimental research. And I think PIs should start to care more about raw data. I mean look at a lab staff with 8 undergraduates, 1 post doc, 1 lab manager as scientific staff and each one presumably has to work for high impact publications. So I ask you, dear science community, were is the control and quality of good experimental work?

        1. Sorry, I meant 8 graduates. Phd students, i guess. And 3 undergraduates. I just think that good institutions should invest more in higher qualified scientific staff and technicians. In some institutions I see way to much Phd-Students being employed, just because they are “cheap” and can be put under pressure.

          1. That is a perfectly valid point. This happens all over the world. Unfortunately this jobs, for trained technicians hardly exist! In Brazil it is impossible to get them, you can get short term fellowships but not the position!

  6. Major points for the PI on this. I sincerely wish she hadn’t added the “but we’ve done it right now and really it works”. I hate it when they do that – retracted is retracted. That should be saved for real peer reviewed papers.

    1. Here, I view, “we believe the study and will be submitting a corrected version shortly” as different from “we still believe the study period” (where no new paper is forthcoming). It’s a timing problem. It would be nice to say, “we did it again and it works, see J Exp Med 2013 xxxxx.” Presumably Cell didn’t want to wait, nor should they. The key here will be to look at Dr. Ting’s publications a year from now and see whether the study really was republished, and if so, where.

      1. Agreed, in general. The method is interesting, and in fact I want to believe. It’s just a matter of personal preference here. I feel that saying “this is retracted, but we did it and it works” it’s extraneous information, sort of making an excuse for the retraction.

        Plus, Cell didn’t have to wait at all for some sort of indication that the approach works. The point is to retract and say why. Future fraud-free publications will certainly stand on their own, particularly for an approach this promising.

  7. Dr. Thyagarajan has vanished from the Clark & Elbing website all of a sudden…

    For posterity, from Google’s cache:

    AMAR THYAGARAJAN, PHD / Technology Specialist

    PhD Biological Sciences, State University of New York
    MSc Zoology and Genetics, University of Calcutta, India
    BSc (Honors) Zoology, University of Calcutta, India

    Amar Thyagarajan joined Clark+Elbing in 2012. His work focuses on the drafting and prosecution of patent applications, primarily in the fields of cell and molecular biology, biochemistry, and neuroscience.

    Amar’s research focused on the molecular basis of nervous system development. For his doctoral work, he employed biochemical and molecular techniques to elucidate a post-transcriptional genetic network that controls key aspects of early brain development. As a post-doctoral fellow at MIT, he explored novel strategies to study the molecular mechanisms underlying neuronal synapse development and plasticity. Amar has co-authored an invited review article in Experimental Cell Research and research articles in Journal of Biological Chemistry, Journal of Comparative Neurology, Cell, and Brain Research.

    Select Awards & Honors
    Post-doctoral Fellowship, Autism Speaks
    Distinguished Doctoral Dissertation Award, State University of New York

    [email protected] 617.897.2072

    1. Incredible catch! I don’t know many scientists who work for law firms, but this is an interesting development that should be monitored. Why has Clark+Elbing not provided a response and comment?

  8. The first author is a bit of splicer of gels and blots
    http://www.jbc.org/content/279/48/49680.long

    Having said that he is fairly overt with his splicing and as he isnt lining up series of proteins or quantifications or loading controls, I don’t think this represents misconduct

    As for the last author, if she is engaged in misconduct you would never spot due to the type of work she is involved with. Which is why I always say concentrating on image fraud (surprising useful although it can be) is only catching a small, unrepresentative and at times downright odd proportion of actual misconduct.

    As it is, it seems that half the authors are disagreeing with the other half. And as per ususal the institution seems in no mood to make whatever investigations it did – if any aside from those of the last author – public. So we are expected just to take their word for it.

  9. Now the public knows that PI and the institution are OK. Will the Post-doc be quietly given a nice job in a quiet place? Anyway, it’s the end of story. Wrong.

    Institution must now contact FBI. Post-doc must be charged with fraud based on the existing institution and PI documents and testimony. Then, the Post-doc is facing jail time and heavy fines. Post-doc has no choice, but to start talking… The end of story is when the court decides who is wrong and who is right.

    The beauty in this case is that both institution and PI already declared falsification. Someone did this, or both, or three. After just one court precedent, the fraud in science should be cut by half. You think so?

    No, science is much smarter than you think. After the court precedent, no one will ever say the word falsification, only honest mistakes: just the policy change, self-correcting.

    1. This reminds me of the Amy Wagers case familiar with RW readers:
      http://www.retractionwatch.com/category/by-author/amy-wagers-retractions/

      Anyone wishing to know the reality of science-fraud investigations by institutions should be aware: Blame the post-doc and then Harvard professors promote the PI.

      I would suggest it may be worth science-fraud-hunters take a look at the work of one of the professors involved in Wagers promotion, who is in the same field of research, and thanked in the Wagers retracted paper, David Scadden. Professor Scadden was one of the experts who proofed the now retracted paper:
      http://bloodjournal.hematologylibrary.org/content/112/3/519.full.pdf+html
      page 530.

      1. Ya, but it actually started ab. 1995, with the case of David Baltimore (MIT), when US Rep. John D. Dingell proposed to investigate frauds in science, as any other fraud, with FBI. As one web site says, he “brought in two scientists–Walter Stewart and Ned Feder–and the U.S. Secret Service to probe the Baltimore case.” They did the probe, but MIT did not like it. Scientists made a correction – Baltimore was found innocent, his assistant, Imanishi-Kari, was found guilty. And, at this time it was decided on the now famous self-correcting mechanism. Later, I spoke to Walter Stewart about my case; he said he has children and wants normal life. His web site then closed. Stewart and Feder were honest people and very active. NIH made them sorry they started their activity. THERE ARE NO SUCH PEOPLE NOW. All that is history, but I remember it well.
        And your name?
        I posted some proposal here:
        https://www.retractionwatch.com/transparencyindex/

        1. No, my memory isn’t that good. I checked again. It appears that Imanishi-Kari was finally found not guilty. There were several investigations, also by NIH, the blame was shifted several times.

    1. What’s interesting about the ‘corrected’ paper is that date of submission – Oct 4 2012. Dr Ting doesn’t waste time to score another publication it appears. By the way: another instance of a “retracted work that appears again the literature” to add to the Retraction Watch list.

      1. One possibility is we are looking at a rather unusual variant of the priority dispute.

        The first author is saying that he is looking forward to a fair hearing from the ORI. Lets hope he gets it and lets hope all adjucating bodies (whether ORI or the University) emphasize transparency of process.

      2. I’m not sure I’d call a paper that is so frank about how apparently poorly BLINC works, even in its corrected form, a “score”.

      3. Since Ting’s lab has had up to two years since publishing the erroneous first paper to recognise and then investigate the problem with the original BLINC method, it’s hardly a rushed effort “to score another publication”! Looks more like a careful effort to get to the bottom of a serious problem and to communicate this to the wider community.

        The PLoS ONE paper is rather fascinating ‘though. I’ve never previously seen a paper in which a group presents their systematic efforts in tracking down problems with the work in a previous retracted paper. The way that the new study is presented makes it perfectly appropriate to cite the retracted paper (which they clearly state has been retracted both in the text and the reference list – note that the retracted paper on the Cell site has “RETRACTED” in massive bold red text across each page).

        From a quick perusal it seems that there is no obvious cast iron evidence of a misdemeanour, since the PLoS ONE analysis simply describes several points at which the previous data couldn’t be replicated (i.e. BLINC with the previously reported constructs and protocols work in HEK cells but not neuronal cultures; the previously described plasmid constructs expressed only transiently and was undetectable at DIV5, etc.) They don’t seem to show that the experiments couldn’t possibly have worked as originally reported, although there may well be other evidence (e.g. available to the investigation team) that supports that interpretation. Interesting to consider whether if Dr. Thyagarajan were to contest this, he could be given the opportunity to redo his original experiments and show that in his hands the experimental outcomes were indeed as reported in the original retracted paper!

        1. A Viola (as my musician friend used to say)
          The patent:
          http://www.freepatentsonline.com/y2012/0129159.html

          The patent was filed in October 2011 and if you go deep into the PDF around page 48 you see Example 4 has strong similarities with the PLOS publication – method described as ID-PRIME.

          Maybe there is a reason why an approach would work with lipoic ligase and not biotinylation – thats beyond my area of experience. Or this could be a conflict between a PI and a post-doc turning really nasty.

          1. Good catch LGR. Something is really funny. @Chris: you would not agree. How can you cite a retracted paper?

          2. Ressci do you mean:

            (i) “what’s the procedure for citing a retracted paper?” or

            (ii) “how can one possibly cite a retracted paper which is now supposed not to exist?”

            If it’s the first, you just have to look at Dr. Ting’s PLoS ONE paper linked in ms. post above. That illustrates a perfectly fine way of citing a retracted paper.

            If it’s the second then one should recognize that a retracted paper doesn’t cease to exist. Nobody is going through libraries cutting out retracted articles from bound copies of journals! The papers still exist in on-line “volumes”.

            This example illustrates something quite important about science, namely that just as in one’s personal life, stuff happens and the way we got from point A to point B isn’t always neat and tidy. We can’t magic into non-existence nasty events in the past. Best to face the problems and deal with them maturely. It would be rather silly of Dr. Ting’s group to write the PLoS ONE paper without referring to the retracted paper. It would be a series of apparently pointless experiments describing a procedure that didn’t work without any sort of context. However those experiments were previously published in a manner that indicated that they did work. Other groups may have been interested in the now-retracted work and may have tried stuff as a result. The PLoS ONE paper is a very explicit indication of the efforts of Dr. Ting’s group to identify the nature of non-reproducibility of a previously published study and to communicate this to the wider community (as well as getting their own research efforts back on track presumably). It doesn’t make much sense to do this without referring explicitly to the retracted paper.

            More generally I don’t see a problem with citing retracted work. There may well be good reasons for doing so…

  10. There are several things odd about this PLoS One article. First, it seems to affirm that Dr Thyagarajan is the inventor of the BLINC method because BLINC with the original constructs work at inter-cellular contacts . Second, it is also odd that the PLoS One article solves expression problems by replacing the CMV promoter with a CAG promoter. Let us look at this for a moment.

    The Cell paper used previously published constructs (with CMV promoters) and simply introduced the AP and BirA tags into previously published neuroligin and Neurexin constructs. I think a similar AP-NLG1 construct has been reported in another paper/talk, e.g., http://www.ncbi.nlm.nih.gov/pubmed/16554831, even though they used a different method of transfection and different aged neurons. So it seems that the plasmid backbones are the same and generally work in neuronal cultures.

    The plasmid clearly works in HEK cells, so the construct is fine. The CMV promoter has been used for many years for neuronal expression and even with the nucleofection method in dissociated cultures, for e.g., http://www.jneurosci.org/content/25/45/10469.long (see, for e.g. figure 1 at DIV 12-14). So there seems nothing inherent about nucleofection that stops the CMV promoter from expressing at DIV 12-14.

    Thus, it is odd to see that the CMV plasmid suddenly stops working in neurons in the hands of a few just because it has a CMV promoter. Does this mean that neuroscientists who have been using CMV promoters need to be able to explain their results in light of this new paper? It seems very odd that the reviewers did not catch this.

    The timing of this article and the patent application in 2011 also seem odd.

  11. To follow up on the promoter and expression issue:

    The CAG promoter used here induces quite massive over-expression. This is well known. The Cell paper, in fact, controls for the over-expression by titrating the amount of DNA introduced into neurons. It is well known in the field that over-expression of NLG and NRX causes synaptic changes. The PLoS One paper does not address the effect of the over-expression that the CAG promoter will induce. There must be a good reason why most of the previous NRX and NLG papers used CMV promoters, and discussed in great length these over-expression effects.

    The PLoS One article uses a CAG promoter with 3X-AP peptides, which obviously will give a stronger signal simply because of the two extra AP tags.

    It seems there is more than what meets the eye with regard to this PLoS One paper and the true motive behind the PLoS One paper and the simultaneous retraction of the Cell paper may need to be investigated beyond what the authors claim.

  12. I’m a bit surprised everyone is ignoring the MIT investigation’s findings. Of course those can’t be made public, for legal reasons, but it certainly is not simply one person’s word against another.

    1. I am not ignoring it, I just can’t read it.
      I have yet to hear of a successful defamation action regarding a report by University ethics committee (and fear of litigation does not seemed to have concerned Cell), although I have heard of parties taking out injunctions to prevent publication.
      In general ethics committees operate under qualified privilege, in so much a court would be highly unlikely to find against them in defamation proceedings unless it could be proved they had published findings they knew were false at the time.

      In other words MIT would not need to prove their findings were true, simply that they were not made in malice. Whereas the complainant would not only have to prove that the findings were in error, but would also have to demonstrate that MIT had been acting in bad faith.

      BTW it is qualified privilege that also provides (some) protection for whistle blowers – although they can step outside that protection. You can read more about this issue here
      http://ebm.rsmjournals.com/content/224/4/220.full

      Ref 63 is probably the relevant one here. A claim for defamation by Shen vs the University of Minnesota being dismissed on summary judgement.
      Although Ref 61 is also an interesting one

    2. I don’t think people are completely ignoring the investigation but perhaps not embracing it fully either (the little that we know). Two recent cases on Retraction Watch, Sebastiao and Elton, illustrate the point. In the former, a journal seemed to ignore the findings of the University for unknown reasons and in the latter, ORI questioned the original findings and a subsequent investigation came to a different conclusion. As many others have pointed out here, the University investigating itself is a bit of a problem. There is clearly the appearance of conflict of interest. Whether this impacts investigations or not likely depends on the particular institution and situation.

  13. I think it’s a little puzzling that we’re trying to pursue some sort of conspiracy here. Usually the simplest explanations for a phenomena are also most likely to be true. It seems a little too drawn out to spend this much time and effort (of postdocs, grad students, technicians) to invent a plot to disenfranchise another individual, and if you were a postdoc/grad student you are clearly not making the best use of your time.

    An institution found the post-doc responsible for falsification and fraud, which is much more serious than just data being irreproducible.The PLOS1 paper only goes as far as explaining why the Cell paper is irreproducible using the constructs from it. Clearly there must have been some additional evidence for an institution to declare falsification and fraud than just an explanation for the paper’s irreproducibility, otherwise the institution is opening itself to being sued for overstating its claim.

    1. Sam stated: “If you were to review all of the federal register notices coming out of ORI you will find that PI’s or supervisors are found culpable only if they were aware of the falsification (or directed it) even if they themselves did not have a direct hand on the data. Being a bad mentor or boss is not misconduct as far as law is concerned.”

      Unfortunately, university investigation committees do NOT always follow Sam’s logic and ORI’s “law” – I have consulted for several professors on such cases (where the falsification was done by a postdoc or student):
      In one case, the professor WAS found guilty of misconduct by the committee for “being a bad mentor” — after the professor asked to be investigated when the committee “waived” on the student! In another two cases,
      the professors were investigated for one to three years and held “responsible for the falsifications” that were
      done by a postdoc or student, as an issue of “reckless misconduct.” This is grossly unfair to them. I do wish
      the university officials would watch carefully over such investigation committees and follow “the law” as Sam describes.

      1. What should be done in cases where it has been determined that the mentor does bear some responsibility ?,i believe the question to be asked is, Does this particular incident in any way, reflect on the quality of the training in that lab ? For example, is there an unrealistic ratio of trainees to mentor ?.

        I believe that where ORI has established that a mentor does indeed bear some responsibility for a trainee’s misconduct,the agency funding the training grant should be notified with a copy to the mentor’s institution.
        It should then raise a question regarding the quality of training at the time of the grant’s renewal.
        Perhaps, funding agencies could establish reasonable ratios of trainees to trainer to assure adequate supervision.

        Don Kornfeld .

        1. I think you’ll find numerous examples at ORI and on RW where PIs are sanctioned. Sanctions can range from a mandate for extra institutional supervision to repayment to disbarment from receiving federal funds for 2 years to permanently.

  14. If I recall, the current Director of the US National Institutes of Health, Dr. Collins, was involved in a similar type of retraction (graduate or postdoc misconduct) many years ago. Maybe Ivan and Adam can dig this example up and present it as some sort of “historic case”. It might be interesting to revisit some of the older cases from time to time to gain perspective on our current problems.

    1. See this ORI case summary at Pages 26-27 http://ori.hhs.gov/images/ddblock/ori_annual_report_1997.pdf

      “Amitav Hajra, University of Michigan (UM): Based upon a report from UM, information obtained by the ORI during its oversight review, and Mr. Hajra’s own admission, ORI found that Mr. Hajra, a former UM graduate student, engaged in scientific misconduct by falsifying and fabricating research data in five published research papers, two published review articles, one submitted but unpublished paper, in his doctoral dissertation, and in a submission to the GenBank data base. Mr.Hajra’s doctoral training and research was supported by PHS grants, and his experiments were conducted at NIH’s National Center for Human Genome Research (NCHGR). Mr. Hajra began his graduate research at the University of Michigan with Dr. Francis Collins as his mentor. When Dr. Collins later accepted the position of director of the NCHGR and established a research laboratory at the NIH, Mr. Hajra continued his research on the NIH campus. The possibility that data had been fabricated or falsified first came to the attention of Dr. Collins when an editor informed him that reviewers of a manuscript had questioned the authenticity of a figure. When intervening events and a survey of laboratory notebooks and other data confirmed deep concerns, Dr. Collins confronted the
      student who admitted to fabricating major portions of his dissertation research and related research publications. The UM, NIH and ORI were notified. Dr. Collins also submitted retractions and corrections of the relevant publications and databases. ORI asked the UM, where Mr. Hajra was completing his final year of medical school, to conduct a formal investigation. The following research reports (1-5) and review articles
      (6-7) contained falsified and fabricated data. . . .Mr. Hajra was found to be solely responsible for the data
      falsification and fabrication and no patients were involved in the research. Mr. Hajra accepted the ORI finding
      and entered into a Voluntary Exclusion Agreement in which he voluntarily agreed, for the 4-year period
      beginning July 7, 1997, to exclude himself from any Federal grants, contracts or cooperative agreements and to exclude himself from serving in any advisory capacity to the PHS.”

  15. I do not understand why Cell would not contact the first author regarding the retraction. I assume this is atypical, given this previous editorial from Nature: http://www.nature.com/nm/journal/v9/n9/full/nm0903-1093.html.
    Especially here, given that there are only two authors, why would they not have contacted everybody associated with the article? Nature claims they try to resolve such an issue prior to retraction, and such retractions consequently, take a long time to sort out. In light of that, this situation seems odd.

      1. @StrongDreams:
        From RetractionWatch above:
        “Update, 1:30 p.m. Eastern, 2/15/13: Thyagarajan sent the following comment:
        ‘I was not contacted by Cell about their decision to retract the paper. I want to be clear that the retraction was done over my objection.’ “

        1. How odd. I suppose Cell took Dr. Ting’s word for it that he could not be reached or declined to participate. Although, that’s what you would expect a faker to say. I guess the patent conspiracist’s case just got stronger.

          1. It could be a matter of ownership.

            First study – now defunked – who owns the work and IP?

            Second, new improved study – who owns the work and IP?

            Could this simply be about ownership and potential Gold?

          2. @Stewart,
            See the above post by LittleGreyRabbit. Dr. Thyagarajan’s technique “BLINC” is not patented. The new technique, “ID-PRIME” is patent-applied-for in 2007 and revised in 2011; Dr. Thyagarajan is not listed as an inventor.

  16. @StrongDreams,
    Based on the first author’s statement if Cell did not contact him with a notice of retraction in which they informed him of their decision to retract how is he supposed to sign?
    @Stewart,
    You are correct – there may be a dispute over the inventorship on the patent application. The Cell paper was most likely very strong prior art against this patent application. Now that it’s gone, the patent application can sail through.
    @James,
    To address your statement “it seems a little too drawn out…,” well that is an environment where grad students and post-docs tend to harm each other. There is nothing new in this idea. Dr Thyagarajan claims his data were tampered/deleted. For a moment, let me believe he is wrong. I have read Dr Thaygarajan’s Cell paper and now, the new – PLoS One – as well. I am very, very intrigued by the most total similarity of content. I still find the idea of swapping a standardly used promoter like CMV with a CAG promoter to make experiments work, odd. In the absence of the retraction, such a swap would have elicited a nod, at best.

    And in this (comparative) light, I must admit that Dr Thyagarajan could well be right. However, the more important question that should concern us is: what is going on? If my reading of Hamlet is not wrong, methinks “there is something rotten in the Ting Lab”. Therefore, if someone did tamper/delete Dr Thyagarajan’s data to make it look as if he falsified it, I hope MIT has investigated this thoroughly and proven that his data were not tampered with and no data were deleted. However, if the investigation proceedings will not be made public, as indicated by Littlegreyrabbit, then we have a situation that does not bode well for science, the researchers and academia at large; then who is to know what really happened in that lab, what is the real motivation behind this retraction and the simultaneous publication of the PLos One paper?

    1. That is absolutely correct. The investigations and retraction notices are cut at the most important moment. The result is that we almost never know what the other side says or what really happened. And the culprit is secrecy. That’s why I believe that there should be a web site that will publish the stories from personae dramatis. Everyone has right to tell his story and publish documents in his possession. All that confidentiality is at the level below the level where one actually wants to publish all this.

    2. Few points/ questions:

      1. Wish the minutes for the meetings of the investigation committee could be available.
      2. Wondering what the percentage of cases where the university employed investigating committee found the alleged researcher not guilty.
      3. Is there any criteria to select the investigation committee members to avoid (i) peers pressure, (ii) mistakes in judgement due to ignorance in the field of research under investigation, (iii) tendencies to run the investigation in favor of the whistle-blower ?
      4. Is there any rule to retract a paper from a highest impact journal ? or, it is solely in discretion of the PI ?
      5. Anybody can bring allegation against his/her students/ postdocs. University sets up an internal investigation committee and they find the truth and give the verdicts. How many students/post-docs have financial back-up to go to ORI to further appeal to prove innocence and being victimized?

      1. Kris, your scenario about students and postdocs who might “go to ORI to further appeal to prove innocence” is not feasible. A decision by ORI “not to pursue a university finding of misconduct” is just that — there is no public statement made by ORI that the university misconduct conclusion was “wrong.” [Since 2007, after my departure, ORI stopped publishing summaries (without names) of its cases with no ORI-findngs of research misconduct in its Annual Reports.] ORI does not state the person in such a case is “innocent” or “cleared” by ORI. ORI does not publically “overturn” university conclusions. Students or postdocs who want to challenge a university investigation committee’s conclusion would have to appeal to university officers or trustees or to a court.

        1. Hi Alan,

          Thanks for your clarification and there you go. I know who I am talking to and I respect your experience in this field. However, I am not aware of any cases (without minor exceptions) where university led investigating committee found the allegation wrong and delivered judgment against the PI although almost everybody knows and it is a matter of fact that PI-Postdoc relationship at the very high-end of research is not always smooth, especially if the postdoc is working to make a bright career in the same line of research and there is a possibility of competition. Further, universities always back their faculties because it is the question of their reputation, even if it leads to one or two article retractions. I don’t think any institution would do a fair judgement for the sake of a postdoc’s career against the reputation of their well-cited faculty.

          I rephrase my question here: How many students/post-docs have financial back-up to go to court to further appeal to prove innocence and being victimized if the university officers or trustees simply ignore their supporting data and related proofs?

          Most importantly and specially for postdocs/students from overseas who are here with a visa, they can’t stay here to fight this battle because their visa expires once the university announces their decision and cancels their registration. Do you have any suggestion for them on how to handle this situation ? I bet, everybody would say to maintain the work-ethics and avoid this situation. I would say, same is expected from the both ends.

        2. In your experience, has anyone actually gone to court and won fairly? I mean this is MIT we are talking about and one of their self appointed “celebrity” for lack of a better term. MIT has made Dr Ting a local science hero. I never expected them to treat this case fairly and how could they? They went with what she had to say and gave their verdict. Will the court of law do the same? side with a world renowned institution and their PI against a no name post doc? Where is the justice? what does this say about future scientists willing to come to MIT.

        3. Hypothetically he could contact the ORI with – not a request to be cleared – but with a charge of research misconduct against Professor Ting for
          a. a deliberately misleading and falsified publication in PLOS
          b. misusing the retraction process for a paper she knew was valid

          One of these is definitely misconduct under the ORI definition, I think both are. Although I admit the ORI might be unwilling to take it up, or if they did feeling unable to come to a position of absolutely certainty which seems to be the criteria before they issue a finding. After all, most of their findings are in consensus with the accused person – and Prof Ting has no incentive to agree to anything.

          An interesting thought: an ability to replicate work is generally not seen as proving misconduct – in this instance would an ability to replicate the original paper be seen as proving misconduct in the subsequent publication?

          I wonder if the ORI would be able to achieve more if they had true police powers – such as the power to seize email communications etc of work email accounts and phone texts etc. We have had a couple of fascinating cases recently in Australian politics when all kinds of things are revealed when that happens.

    3. > I still find the idea of swapping a standardly used promoter like CMV
      > with a CAG promoter to make experiments work, odd.

      There are all sorts of reasons that this might work, depending on the nature of the construct.
      For example, there may be secondary structure or emergent sequences.
      You’d need to dig into the sequence to find out, and even then there are enough things
      unknown about cellular context interactions that you won’t necessarily be able to tell anything.

  17. There is the old saying from Harry Truman, “the buck stops here”. This political phrase should also apply to all faculty and laboratory heads. Ting is fully responsible for this mess. She was PI on the study and the corresponding author on the publication- the buck stops with her. A careful PI who who cares about analyzing data instead of being in the limelight should be the biggest skeptic of results. She should have been able to critically assess and interpret the data in their raw and polished forms. There are students and post-docs who only yield “positive” results and a skilled PI should be wary of these individuals, since every scientist knows that most experiments yield “negative” or ambiguous data. My guess is that other members of her lab had their suspicions and did not speak up until after publication. Nonetheless, Ting is the PI and even though there was a bad seed in her group, the buck does stop with her.

    1. Agreed. To quote myself from an earlier post: I believe that where ORI has established that a mentor does indeed bear some responsibility for a trainee’s misconduct, the agency funding the training grant and the mentor’s institution should be notified.
      When the training grant comes up for renewal,such reports should raise a question regarding the quality of their training. Such a policy should motivate institutions to demonstrate that appropriate steps have been taken to improve the quality of its training program., eg,, fewer trainees per mentor.

      Don K

  18. I would blame Ting for not knowing what was happening in her lab. Some PIs just start dreaming of a publication in science or nature after seeing some exciting results from the lab members without knowing the authenticity of it.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.