Retraction Watch

Tracking retractions as a window into the scientific process

Leading cancer vaccines researcher retracts paper for figure “discrepancies” flagged by watchdog blog

with 23 comments

Gerold Schuler, a German immunology researcher who shared the 2006 Deutscher Krebspreis — aka the German Cancer Prize — for his work that contributed to cancer vaccines has retracted a paper in International Immunology following concerns raised by a German science watchdog blog.

Here’s the notice:

After publication of our article “Optimizing the exogenous antigen loading of monocyte-derived dendritic cells”, several discrepancies have been brought to our attention. In Figure 1 and Figure 2A, in contrast to what is stated in the figure legends, identical control staining is shown within the various parts of the figures. In Figure 3, the dot-plots with values shown as 10.2 and 10.8 were obtained using the same data file, as were those with 12.9 and 14.1. In Figure 7, the dot-plots with values of 9.8 and 16.2 turned out to be derived from the same data set. The appropriate experiments are being repeated, but in the meantime the authors would like to retract the original manuscript and apologize for these errors.

The paper has been cited 14 times, according to Thomson Scientific’s Web of Knowledge, most recently by a PLoS ONE paper, “A Double-Blind Randomized Phase I Clinical Trial Targeting ALVAC-HIV Vaccine to Human Dendritic Cells.”

The paper’s use of the same controls in a number of images — which has tripped up a number of other authors about whom we’ve written — was noted last year here, here, and here by the Abnormal Science blog. The posts also raise questions about other papers by Schuler’s team.

Joerg Zwirner, the author of Abnormal Science, writes in that first post that he told senior author Gerold Schuler about the “severe irregularities” in the article last year. He also notes how prominent Schuler is:

In 2005, when the paper was published, the first and corresponding author Dr. Detlef Dieckmann left academics to become a dermatologist in private practice. Therefore, I contacted the senior author Prof. Schuler.

Since 1995, Gerold Schuler has been director of the Department of Dermatology at the university clinics Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg.
In 2006, he was awarded the German Cancer Prize for his work on dendritic cells.
He is speaker of the Collaborative Research Centre (Sonderforschungsbereich) 643: Strategies of cellular immune intervention.
At his clinic and elsewhere, melanoma patients have been treated using dendritic cell-based cancer vaccines, with very limited clinical efficacy so far (see here).
On July 7, after two terms, he retired by rotation from the Senate as well as the Executive Committee of the German Research Council (DFG), Germany’s largest research funding organisation.
From 2005-2011, he was also a member of the DFG Committee of Inquiry on Allegations of Scientific Misconduct.

What’s unclear, of course, is how the discrepancies ended up in the now-retracted paper, and why no one noticed before the study was published that some of the dot-plots “turned out to be derived from the same data set.”

We’ve contacted Schuler and the editor of International Immunity for comment, and will update with anything we hear back.

Update, 3 p.m. Eastern, 5/30/12: We received this message on behalf of Tadamitsu Kishimoto, the journal’s editor in chief:

The issues raised by the “Abnormal Science” blogs in relation to the original publication were apparently important in triggering the investigation by a University Committee for Good Scientific Practice at Erlangen-Nuremberg University, which also had input from an external expert.

No other Retractions are planned at the moment.

Update, 2 p.m. Eastern, 5/31/12: Schuler responded:

The errors were reported to me via by an e-mail of Dr. Jörg Zwirner to which I immediately responded, indicating that I will inform the University and ask for an independent investigation by the respective committee.
Unfortunately, Dr. Zwirner at the same time started what I consider a smear campaign targeted at me personally e.g. by writing e-mails to employees of a large number of Universities without waiting for the results of an independent investigation by the FAU (Head: Prof. Dr. Karl-Dieter Grüske).
There are no other retractions planned.
Comments
  • Clare Francis May 22, 2012 at 10:19 am

    One thing you can say about Gerold Schuler is that he does have comedic qualities, intentional, or not, I am not sure.

    On German national radio, Deutschlandfunk (funk is a spark)

    http://www.dradio.de/dlf/sendungen/forschak/1569019/

    in the wake of the Nobel prize being awarded to Ralph Steinman shortly after his deth in 2011

    Gerold did compare himself, and others to Jesus’ disciples.

    “Wir waren ja wie die Jünger Jesu”
    Mediziner aus Erlangen über die Arbeit mit dem verstorbenen Nobelpreisträger Ralph Steinman
    Gerold Schuler im Gespräch mit Christiane Knoll

    On a more serious note I have concerns about another of the papers by the same group:

    J Immunol. 2002 Jul 15;169(2):1110-8.
    TCR-like human antibodies expressed on human CTLs mediate antibody affinity-dependent cytolytic activity.
    Chames P, Willemsen RA, Rojas G, Dieckmann D, Rem L, Schuler G, Bolhuis RL, Hoogenboom HR.

    The paper looks like it contains decorations, rather than a record of what went on
    and closely related to a collage.

    My analysis is from what I can see in the publication.

    Figure 2.
    http://www.jimmunol.org/content/169/2/1110/F2.expansion.html
    Figure 2, panel A has not been created by a machine, but by a human hand. Notice the way the dotted line, or rather the stretches of the line wobble. The same hand is at work later in the publication.

    Figure 4.
    http://www.jimmunol.org/content/169/2/1110/F4.expansion.html

    What is going on in figure 4, the middle panel ( i.e. on the right, halfway down)?
    Sometimes the parts of the dotted line which are marked curve back on themselves.
    Larger version.
    http://www.jimmunol.org/content/169/2/1110/F4.large.jpg

    Figure 5.
    http://www.jimmunol.org/content/169/2/1110/F5.expansion.html
    The “tranduced/non-enriched” outlines look the same in both panels to me. You think that this line is different in the lower panel because the control is a bit different so there is no gap between them.

    Figure 6.
    http://www.jimmunol.org/content/169/2/1110/F5.expansion.html
    There are no data points. We are just told “with SDs not exceeding 10% of mean values.
    Data from one representative experiment are shown”.
    Larger version.
    http://www.jimmunol.org/content/169/2/1110/F6.large.jpg

    Notice the way that the dotted line in the bottom 2 panels wobbles just the way the dotted line wobbles in figure 2, especially within in stretch of line drawn. You can also see this in the top left hand panel

    • Pinko Punko May 22, 2012 at 9:45 pm

      CF- are you familiar with Biacore data? Your first link looks similar to what Biacore data looks like. You should compare those images to Biacore responses, try a Google image search.

      • Clare Francis May 23, 2012 at 3:05 am

        There are multipe things that caught my attention.

        If papers start looking like computer screens only accessible to the few then that is their problem. When “aspects of reality”, an essence of science, get too far away from reality people need to reconsider what they are doing. Many Journal of Immunology papers (and others) look like walls of text. We are just left “peering in”. I do wonder if much of what we call “science” is simply a front for the manufacturers of the next, more expensive piece of machinery. They are machines as they help the manufacturers. To get something published you need something from a machine in list A, and something from a machine in list B.

        Recently on this blog I saw somebody write something about “keep science small”. I wonder if that means science you can do in your head, or in a few heads?

        I do not think that it Luddite, but whether it was pressing button B on a Biacore machine, or not, does not make it science. It is measuring.

        Even if it were Luddite the point is that the technology is not the problem, but the failure to recognize that technology, or combinations of technologiey, do not make science.

        Here endeth the lesson:

    • Sebastian May 23, 2012 at 5:19 am

      “Deutschlandfunk” means “German Broadcasting Station” – not “German Spark”.

      He wasn’t comparing himself to Jesus’ disciples, but stated that they were _like_ the disciples, in regards to their belief in the newly developed theories regarding immunity which at that time were not accepted by the mainstream (which sounds a lot like that funky story in parts of the Bible…).

      Please, if you want to crucify someone (pun intended) because of this retraction, do make an effort to understand what you’re reading. If you’re not fluent in another language, then do not use another person’s statements made in that language for making accusations.

      Also, as stated by Pinko Punko, if you’re not familiar with the data output of a particular piece of software, then try to familiarize yourself with it.

      • Clare Francis May 23, 2012 at 6:44 am

        Funk is spark, not funk as English speakers might understand.
        It is not “funky”.

        “Like” is comparing. He was comparing himself with Jesus’ disciples.

      • Clare Francis May 23, 2012 at 6:49 am

        Familiarity with the emperor’s clothes, sorry I mean “data output”, was not the main point.
        As I wrote there are multiple things. If “data output” is odd it is still odd. “Data output” still needs to make sense.

      • Sebastian May 23, 2012 at 8:46 am

        Dear Clare,

        I happen to be German, so please trust me that “Deutschlandfunk” does not mean “Germany spark”. What you are referring to is “der Funke” (please translate.google it).

        Familiarity with analysis output is certainly important – I will not tell my colleague who is a neurosurgeon how to interpret MRI scans or how to plan a surgery based on these data. Conversely, he will not tell me how to handle quality issues in sequencing data, or alignment of reads to low-complexity regions.

      • blatnoi May 25, 2012 at 4:39 am

        He’s right, wot wot. I ain’t one of ’em Germans, but I speak their dodgy language fluently. Hands down for Seb on the ‘Deutschlandfunk’ charge, and while you might argue about the semantics of similies for the Jesus thing, I’ve got to give the edge to Seb. The bigger point is that you should still be careful when relying on Google translations. German is one of the best but still it sounds unnatural translated and you get these big errors from time to time, but the thing really butchers the other languages that I know. With this great increase in computing power and software engineers, we still don’t have perfect translation software. Once we start getting perfect translations of languages by some AI algorithm, then I’ll start worrying about machines taking over and enslaving us.

  • Ressci Integrity May 22, 2012 at 10:39 am

    As early as 1997 the investigations started on some of the german investigators. Here is an example:
    Nature 387, 750 (19 June 1997) | and in 2000
    Abbott A: German fraud inquiry casts a wider net of suspicion…Nature 405, 871-872 (22 June 2000) | doi:10.1038/35016207

    Few names included in this report are currently directors of institutes…

  • Dave May 22, 2012 at 11:23 am

    Unless I am mistaken, Dieckmann is the corresponding author on this paper, so some caution should be exercised when going after Schuler. The other posts on Abnormal Science about some of Schuler’s other papers are not terribly convincing in my opinion, at least right now.

    • Fernando Pessoa May 22, 2012 at 4:19 pm

      http://www.icmje.org/ethical_1author.html

      “An author must take responsibility for at least one component of the work, should be able to identify who is responsible for each other component, and should ideally be confident in their co-authors’ ability and integrity”.

      O.K. that is “aspirational”, but an author should really know what has been going on.

  • Ginsberg May 22, 2012 at 1:27 pm

    Please elaborate Dave. Only the corresponding author is accountable for the authenticity of the paper?

    • Dave May 22, 2012 at 3:54 pm

      No, of course not, but the corresponding author is ultimately responsible for the contents of the paper. This has been emphasized again and again and again on this site. When you submit a paper these days you even have to sign off on this statement if you are the senior author. All I am saying is that some caution should be used before jumping to wild conclusions.

  • Mr. M May 22, 2012 at 3:27 pm

    This Abnormal Science Blog, however, seems to be dead for a while. It is such a great pity for it is a wonderful blog.

  • amw May 22, 2012 at 3:33 pm

    I agree that the figures in the J. Immunol 2002 paper are odd in several ways – dotted lines that wobble backwards, strange line graphs with no data points. The legends either miss important information or are simply incorrect – for example saying the traces in Figure 5 are histograms.

    However, producing strange figures is not misconduct. Some of the blame goes to the reviewing process – any referee spending more than a few minutes on the paper would have noticed at least some of these issues – so essentially it is reasonable to conclude that the paper was effectively not refereed.

    However the Figure 5 transduced / non-enriched traces are the same, which is duplication. The authors have already admitted that this happened in the Int Immunol paper. If one paper is found to have been produced by irregular processes, then it is reasonable to assume others were also affected – possibly many more.

  • Clare Francis May 22, 2012 at 3:43 pm

    In reply to amw May 22, 2012 at 3:33 pm
    I thought I was losing my marbles.

  • Clare Francis May 23, 2012 at 9:52 am

    In reply to Sebastian May 23, 2012 at 8:46 am

    The point about Funk or Funke is not central. Deuschlandfunk is the national radio station.

    The issue is that we are not sure about the data in J Immunol. 2002 Jul 15;169(2):1110-8.

    There is the issue of duplication in figure 5 (transduced / non-enriched traces).
    The graphs in figure 6 look idealised and not well explained.

    • Sebastian May 23, 2012 at 10:38 am

      So, and in how far is a radio interview of any relevance to your critique?

      The point gets central when someone is misquoting or mistranslating a person’s statements and uses these in a personal attack (“comedic qualities”)…

      Why not simply stick to the science, then?

      • Clare Francis May 23, 2012 at 1:11 pm

        It was not misquoting what he said.

        http://www.dradio.de/dlf/sendungen/forschak/1569019/

        about 2 thirds of the way down:

        “Weil wir waren ja natürlich wie die Jünger Jesu. Das war damals ein relativ kleiner Kreis und wir waren eine eingeschworene Gemeinschaft und wir haben natürlich an das ganz fest geglaubt, an dieses Konzept.”

        equals

        “Because were were naturally like the disciples of Jesus. There was at that time a realtively small circle (group/band) and we were a dedicated/confirmed community and we naturally beleived this concept quite firmly”.

        He does make comparisons. He does not say they were the disciples of Jesus, but he does make the parallel. His words, not mine.

    • Pinko Punko May 23, 2012 at 10:41 am

      CF, the good practice of science is essential. The issue is when things are noted or things asserted that are not appropriate or wrong, given that reputations are at stake. Wouldn’t you rather have false negatives (things you miss) rather than false positives (things you wrongly accuse)? It is clear that you are right a good amount of the time.

      • Clare Francis May 23, 2012 at 1:17 pm

        In reply to Pinko Punko May 23, 2012 at 10:41 am
        You cannot have a perfect world, just try to minimize what is incorrect.

        They are not in the witness box of a criminal court, where the bar is set at “beyond all reasonable doubt”, not even in the witness box of a civil court, where the bar is set much lower “on the balance of probabilities”, but in fact a scientific piece of work where there is a “reverse burden of proof”, i.e. they have to prove something is ture, not for others ot prove it is incorrect.

        My original comments on figures 2 through 6 still stand. To biacore or not to biacore is not the question.

  • Clare Francis May 25, 2012 at 8:39 am

    In reply to blatnoi May 25, 2012 at 4:39 am

    I did not write that Deutschlandfunk means Germany spark.

    What I wrote was:

    “On German national radio, Deutschlandfunk (funk is a spark)”. May 22, 2012 at 10:19 am

    Deutschlandfunk is the national radio station. Remove what is in the brackets. It is not central.
    Listen to the radio interview.
    What’s your translation of what Gerold Schuler said in the interview?

    It is interesting what the root of Funk is, but it is not central.

  • Clare Francis May 25, 2012 at 9:46 am

    Sorry for being Germanic about this, but I think that my original comment was not off beam.

    http://en.wikipedia.org/wiki/Spark-gap_transmitter

    German verb “funken”, literally, “to spark”, also means “to send a radio message/signal”.

    The -funk element of Deuschlandfunk, the verbe funken, and the noun Funke, are all related.

  • Post a comment

    Threaded commenting powered by interconnect/it code.