Retraction Watch

Tracking retractions as a window into the scientific process

Glasgow’s Beatson Institute investigating circumstances of Cell retraction for inappropriate images

with 30 comments

The Beatson Institute for Cancer Research in Glasgow, Scotland is looking into how inappropriate images ended up in a Cell paper that has just been retracted.

Here’s the notice for the paper by Lynne Marshall, Niall S. Kenneth, and Robert J. White:

This article has been retracted: please see Elsevier Policy on Article Withdrawal (http://www.elsevier.com/locate/withdrawalpolicy).

This article has been retracted at the request of the Authors.

Our paper reported that overexpression of tRNAiMet can promote cell proliferation and transformation. It has now become clear that images in several figures are duplications that do not represent the experiments described in the legends. Although we remain convinced that tRNAiMet can have proliferative and oncogenic effects based on subsequent work in our lab and others, given these issues with the validity of the figures, we believe the most appropriate course of action is to retract the paper. We deeply regret this circumstance and apologize to the community. The first author, L.M., has not agreed to this Retraction.

We wanted to know whether there had been an investigation, and why Marshall had declined to sign the notice. Professor Karen Vousden, director of the Beatson Institute, tells Retraction Watch:

We launched an investigation into this matter and we’re unable to comment further until the investigation is complete. All authors were invited to sign the retraction notice.

Senior author Robert White told us he is out on sick leave.

We’ve tried to reach Marshall, now until July a post-doctoral fellow at the University of Alberta in Calgary, for comment, and will update with anything we learn.

The Cell paper has been cited 84 times, according to Thomson Scientific’s Web of Knowledge.

Second-to-last paragraph updated 5 p.m. Eastern, 9/10/12, with new information about when Marshall’s Calgary postdoc ended.

Comments
  • omnologos September 5, 2012 at 2:00 pm

    “Inappropriate images”…I was hoping for at least an R-rated issue of Cell

  • chirality September 5, 2012 at 2:27 pm

    I wonder if this paper formed the basis of LM’s PhD thesis. It would explain the reluctance to agree to the retraction.

  • WHY September 5, 2012 at 3:51 pm

    Not University of Albetra.
    University of Calgary…

    • ivanoransky September 5, 2012 at 3:54 pm

      Indeed. Fixed, thanks.

  • Damien September 6, 2012 at 10:29 am

    I wonder what will happen to a new paper by Marshall et al. (2012):
    http://www.nature.com/emboj/journal/v31/n8/full/emboj201233a.html

    The retracted article is cited several times and uses as a support for conclusive remarks.

  • littlegreyrabbit September 6, 2012 at 5:22 pm

    So reusing Western blots or reusing RT-PCR bands?

    I am very bad at spotting these.

    • Steven McKinney September 13, 2012 at 2:41 pm

      Similar bands:

      Figure 1 C RT-PCR band – second from bottom (label is obscured by RETRACTED lettering, but “RNA” is visible under the first “E” of RETRACTED).

      Page 84 Figure 5? Figure legend label is hidden under RETRACTED lettering. Figure 5 A second last band “DHFR”.

      These two RT-PCR bands are the same image, slightly darker in Fig 1 than in Fig 5, but the lighter smudges in the middle of the bands in the fourth and fifth lanes (counting from the left) are too identical for comfort.

      Another example:

      Fig 1 C top two bands labeled 5S rRNA and tRNATyr are the same image, slightly stretched in the 5S rRNA band relative to the tRNATyr band.

      and another:

      Fig 1 B (immunoblot) bottom two lanes TBP and Actin – images are too close – TBP band is slightly stretched compared to the Actin band.

      The above are proposed candidates for further analysis of image duplication by image experts.

      • littlegreyrabbit September 13, 2012 at 11:53 pm

        Boy, well I looked at Figures 1 and 5 and I can’t detect anything to disprove what you suggest. Ditto with Figure 1B. Ditto with Figure 1C – would you not expect the 5S rRNA to be much stronger than the transfer RNAs (not having worked with these components of RNA machinery)? Some people have good eyes.

        I have to say I would hate to make such a suggestion based on only what is online, so I assume there must be some back story that enabled these accusations to have been made with such confidence.

        My suggestion would be that, remembering that if you are using photoshop to falsify data you are doing it wrong, the best way to avoid detection would be to physical alter the bands a bit with photoshop, print out this altered image and then rescan that print-out so as to avoid any nasty photoshop artifacts.

  • Bonzo September 15, 2012 at 4:40 pm

    See the erratum for Mol Cell. 2012 Feb 24;45(4):541-52. Epub 2012 Jan 25.
    (Molecular Cell 45, 541552; Published online January 26, 2012)
    ” In the above article, inadvertent errors were included in Figures 5E and 6D. In Figure 5E, one of the original FACS profiles was a duplication of another. In Figure 6D, the second panel was a shorter exposure of the top panel. We have reproduced the experiments, and new versions of the figures are included below. None of the conclusions of this article are altered. We apologize for these errors.”

    • Bonzo September 21, 2012 at 12:38 pm

      See also erratum for PNAS article…………

      CELL BIOLOGY Correction for “mTOR associates with TFIIIC, is found at tRNA and 5S rRNA genes, and targets their repressor Maf1,” by Theodoros Kantidakis, Ben A. Ramsbottom, Joanna L. Birch, Sarah N. Dowding, and Robert J. White, which appeared in issue 26, June 29, 2010, of Proc Natl Acad Sci USA (107:11823–11828; first published June 11, 2010; 10.1073/pnas.1005188107).

      The authors note that Fig. 4 appeared incorrectly. The corrected figure and its legend appear below. These errors do not affect the conclusions of the article

  • littlegreyrabbit September 15, 2012 at 7:06 pm

    hmmm

    • poochspotterdownunder September 16, 2012 at 5:41 am

      Its not unusual for a few erratums to come out from a lab after a big retraction like this – as its normal for labs papers to be closely looked at – at least their correcting things publicly.

      • bonzo September 17, 2012 at 10:50 pm

        Or is it because their papers have too many “errors”?

  • Bonzo September 19, 2012 at 10:45 am

    Please see, Nucleic Acids Research, 2008, Vol. 36, No. 11 3757–3764…Figure 1 and 3?

    • kinoko September 30, 2012 at 9:59 pm

      What is suspicious about this one?

      • Bonzo October 4, 2012 at 9:10 pm

        The actin control lanes are the same?

  • Clare Francis October 5, 2012 at 2:27 am

    In replyto Bonzo October 4, 2012 at 9:10 pm

    Dear Bonzo,

    It does not matter.

    ———- Forwarded message ———-
    From: Nucleic Acids Research – Fox
    Date: Fri, Sep 28, 2012 at 2:14 PM
    Subject: RE: concerns Nucleic Acids Research, 2008, Vol. 36, No. 11 3757–3764
    To: “clare francis (clare.francis1946@googlemail.com)”
    Cc: “Barry Stoddard at Nucleic Acids Research (stoddnar@fhcrc.org)”

    Dear Clare,
    Thank you again for your vigilance. We have been informed that all articles authored by Lynne Marshall, including the NAR article (Nucleic Acids Research, 2008, Vol. 36, No. 11 3757–3764) have already been carefully examined by an independent committee at Glasgow University. The conclusions of the committee were that (for this particular article at least) the authors did nothing reprehensible and the publication remains valid. We have accepted the conclusions of the committee and are now considering this issue resolved.
    Yours sincerely
    Keith Fox
    ———————————
    Professor Keith R. Fox
    Senior Executive Editor, Nucleic Acids Research
    Centre for Biological Sciences
    Life Sciences Building 85
    University of Southampton
    Southampton SO17 1BJ, UK
    tel. +44 23 8059 4374
    email foxnar@soton.ac.uk

    From: clare francis[mailto:clare.francis1946@googlemail.com]
    Sent: 19 September 2012 18:13
    To: Fox K.R.
    Cc: stoddnar@fhcrc.org
    Subject: concerns Nucleic Acids Research, 2008, Vol. 36, No. 11 3757–3764
    RE: Nucleic Acids Research, 2008, Vol. 36, No. 11 3757–3764
    Dear Keith,
    I think that figures 1A and 3A share the same actin panels. They also share the same RPC6 panels.
    It is not obvious that figures 1A and 3A represent different experiments. In total there are many panels.
    I think that is unlikely that all the panels in figures 1A and 3A comefrom the same blot/membrane.
    I do not think that the actin panels are the loading controls for the panels above them.
    The bands in the actin panels follow quite a curved trajectory, which is unlike the trajectories of most of the bands in the panels above them.
    C.F.

    • chirality October 5, 2012 at 4:48 am

      Fox should “trust but verify” the findings of the Glasgow committee as it might be independent but not necessarily impartial. The committee might have been reluctant to find misconduct in Marshall’s other papers as it would make here PhD thesis even more bogus than it appears to be. If Marshall forged some blots in the paper, it would take 5 minutes of Fox’s own time to verify it. After all, reputation of his journal is at stake. Fox’s casual attitude can encourage other forgers to flock to Nucleic Acids Research like flies to manure.

      • littlegreyrabbit October 5, 2012 at 10:01 am

        Maybe I am missing something but as I understand the treatments described on Figure 1 and 3 are exactly the same, then there is no reason for them to be different experiments.
        In this case the actin band is showing that the lysates were roughly equal in protein concentrations and then equal volumes of lysates were loaded across however many gels were needed to resolved the various proteins. Whether that satisfies purists or not, is another questions.

        It is not really Lynne Marsh who would be responsible in this instance, as she is not the first author.

      • Fernando Pessoa October 12, 2012 at 3:33 am

        In reply to chirality October 5, 2012 at 4:48

        “Fox should “trust but verify” the findings of the Glasgow committee as it might be independent but not necessarily impartial.”

        As far as I understand the Beatson Institute is part of the University of Glasgow

        http://www.beatson.gla.ac.uk/About/History.html

        “Some of the Cancer Division of the University of Glasgow moved to new laboratory space on site in 1990 and we have now encompassed all basic and clinical cancer research at the Beatson Institute and the University of Glasgow into the Glasgow Centre for Cancer Research.”

  • David Hardman October 5, 2012 at 12:10 pm

    ln reply to littlegreyrabbit October 5, 2012 at 10:01

    Do you think that all the panels in figures 1A and 3A come from the same blot/membrane?

    We do live in the physical world. It is about the reality.

    Let’s think of it as physics. I am not interested in blaming anybody.

    • littlegreyrabbit October 5, 2012 at 2:01 pm

      No (at least I presume not) – the actin panels obviously do, well are duplicates. But if they did load the same amount of lysate on each of the gels, is it that big a deal? They aren’t taking controls from a completely random experiment.

      • David Hardman October 5, 2012 at 2:28 pm

        So the actin panels are not the loading controls for both figures 1A and 3A.
        That paper is not theoretical biology.

        As a general point: how do you know “They aren’t taking controls from a completely random experiment”?

        Not for this group of researchers, but in other groups, I have seen plenty of cases where many bands, not just actin bands, are used for years. In the beginning they are used to represent one thing, as the years go by they are used to represent something quite else. Just because people duplicate does not mean they are not taking controls from a completely random experiment. It is not a perfect world, there is no need to fill in the gaps.

        As people have pointed out on this blog we are presently dealing with incomplete data sets. We do not know how common all sorts of misbehaviour are.

        The issue really is not about actin, even what the best loading controls might be, but about not making assumptions.

  • fittmichael (@fittmichael) October 11, 2012 at 3:43 pm

    this is her facebook, seems to live a happy life at Canada https://www.facebook.com/lynne.marshall.982

  • nofraudplease October 12, 2012 at 12:25 am

    Wow – thanks for this REALLY USEFUL piece of information…..you really are striking a blow for scientific integrity….well done! You must be very proud of your investigative journalism….

  • Ushikawa October 19, 2012 at 2:43 am

    It seems that the retraction notice on the Cell website has been changed slightly, but significantly.

    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 Authors.Our paper reported that overexpression of tRNAiMet can promote cell proliferation and transformation. It has now become clear that images in several figures are duplications that do not represent the experiments described in the legends. Given these issues with the validity of the figures, we believe the most appropriate course of action is to retract the paper. We deeply regret this circumstance and apologize to the community. The first author, L.M., has not agreed to this Retraction.

  • Bonzo December 13, 2012 at 11:46 am

    Does anyone know if there has been any follow up by The Beatson?

    • Beatsonite December 18, 2012 at 6:15 am

      Yes, of course there has.

      • fernando pessoa December 18, 2012 at 10:15 am

        What is it then?

        • scientist February 6, 2013 at 7:22 pm

          Check out the website and the list of researchers

  • Post a comment

    Threaded commenting powered by interconnect/it code.