Retraction Watch

Tracking retractions as a window into the scientific process

Authors in Spain issue string of corrections

with 2 comments

Following a journal probe and questions on PubPeer about their work, authors in Spain have issued four corrections, citing missing raw data for experiments conducted more than 10 years ago.

All papers include the same last two authors, Mireia Duñach at the Autonomous University of Barcelona, and Antonio García de Herreros at the Institut Hospital del Mar d’Investigacions Mèdiques.

Three of the corrections were issued by the Journal of Biological Chemistry, from which the authors retracted three papers earlier this year after a journal investigation concluded they contain reused images, designed to represent different experiments.

Duñach told us the latest corrections are the result of her own initiation:

Regarding the articles in JBC, after knowing that they have been questioned on PubPeer we contacted the JBC editor mentioning this fact and that we did not keep the raw data since they were more than ten years old and we did not have any obligation to keep them, both from JBC or our institutions. We also indicated to the editor that we were willing to repeat all the published results, since they were abolutely reproducible, not only in our hands but also in other people’s experiments. The JBC editor responded that we should repeat several figures in these articles without specifying the concerns in these figures…

She added:

…we repeated the requested experiments and sent the results to the editor that accepted to publish a correction. As we were absolutely certain, these results were totally concordant with those previously published.

The experience has prompted her to change her data-storage policies:

We consider that our responses to the editors have been totally correct, offering us to repeat the results to eliminate any doubt that anyone might have on the validity of the conclusions of these articles. We have also informed the authorities of our institutes of the correspondence maintained with the editors…Finally, we have enforced a new system to store the raw data for longer periods, and to more accurately verify that no mistakes are present in our articles.

Regarding the three articles that JBC withdrew earlier this year, Duñach said:

…they forced us to withdrawn the articles. We [disagreed] with this decision, considering that, according to the [Committee on Publication Ethics] guidelines, retractions should be limited to the articles reporting results that cannot be reproduced and this was not the case.  Moreover, due to the long time after publication, we did not have any way to demonstrate if the mistakes the editor indicated really existed.

First, here’s the correction for “Regulation of E-cadherin/catenin association by tyrosine phosphorylation,” cited 422 times since it was published in 1999:

Several concerns were raised about Figs. 3, 4, and 5 by the journal. Because the original data were no longer available, the experiments shown in these figures were repeated. The repeated experiments are shown and do not affect the results or conclusions of this work. The authors regret the inconvenience these errors may have produced.

PubPeer users have raised questions about the images since December 2015.

Here’s another notice, issued in November:

Several concerns were raised about Figs. 2, 4, and 5. Because the original data were no longer available, the experiments shown in these figures were repeated. The repeated experiments are shown and do not affect the results and conclusions of this work. The authors regret any inconvenience caused by these errors.

beta-Catenin and plakoglobin N- and C-tails determine ligand specificity” has been cited 32 times since it was published by the Journal of Biological Chemistry in 2004, according to Clarivate Analytics’ Web of Science, formerly part of Thomson Reuters. PubPeer users have been discussing the paper since March, 2016.

Here’s the third notice, also issued by the Journal of Biological Chemistry:

Several concerns were raised about Fig. 2, A and C, and Fig. 3B. Because the original data were no longer available, the experiments shown in these figures were repeated. The repeated experiments are shown and do not affect the results or conclusions of this work. The authors regret the inconvenience these errors may have produced.

The transcriptional factor Tcf-4 contains different binding sites for β-catenin and plakoglobin” has been cited 77 times since it was published in 2002. A user on PubPeer raised questions about figures 2 and 3 starting in December, 2015.

After Duñach posted on PubPeer that she had contacted PLOS ONE to correct a paper, that notice appeared in August:

The authors would like to correct Fig 1C, as errors were introduced in the preparation of this figure for publication. The first 6 lanes in the anti-beta actin 24 hours after transfection panel were duplicated in the 48 hours after transfection panel. The authors have repeated the analysis and provide the revised Fig 1 with a new panel C and the underlying data here…The authors would also like to provide additional clarifications for Fig 3. In the published Fig 3, the two bands from panels A and B corresponding to plakoglobin are the same and correspond to the plakoglobin analysis in MEFs PS1 WT in both lanes. The authors decided to present the data from the Western blot split in two panels; in panel A, the differences between PS1 WT and KO and in panel B, the effect of the inhibitor on PS1 WT. The experiments were repeated four times. Panel A was assembled from experiments comparing MEFs KO, control and control plus 1 μM L-685,485; panel B combining those initial experiments (plakoglobin, β-catenin) and experiments involving the control plus two concentrations of L-685,485, 1 and 5 μM.

The authors have replicated the analysis and provide a revised Fig 3 with its underlying images here. The updated Fig 3B includes the three conditions (control and two concentrations of L-685) for all proteins (the 5 μM condition was lacking for plakoglobin and β-catenin in the originally published Fig 3).

Gamma-secretase-dependent and -independent effects of presenilin1 on beta-cateninTcf-4 transcriptional activity” has been cited 14 times since it was published in 2008.

Duñach had this to say about the PLOS ONE correction:

…we responded to an editor message indicating how the two questioned figures have been generated. He suggested us to repeat the two figures again, since the original data was not available either. We reproduced the data and published the correction.

Along with last author Duñach, all papers share the same second-to-last author:  Antonio García de Herreros, who lost a Molecular Cell paper earlier this year (also as second to last author), whose retraction notice cited signs of “intentional misconduct.” In June, we reported that the Institut Hospital del Mar d’Investigacions Mèdiques, his institution, was taking a second look at Garcia de Herreros’s work.

Duñach received an Academia Award from the Catalan Institution for Research and Advanced Studies in 2014.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

Written by Alison McCook

December 30th, 2016 at 9:30 am

Comments
  • Steven Clarke December 30, 2016 at 11:14 am

    Does the Committee on Publications Ethics really limit retractions to only the reason (failure to reproduce) given by the authors in their correspondence with Retraction Watch below?

    Regarding the three articles that JBC withdrew earlier this year, Duñach said:

    …they forced us to withdrawn the articles. We [disagreed] with this decision, considering that, according to the [Committee on Publication Ethics] guidelines, retractions should be limited to the articles reporting results that cannot be reproduced and this was not the case. Moreover, due to the long time after publication, we did not have any way to demonstrate if the mistakes the editor indicated really existed.

  • fernandopessoa December 30, 2016 at 11:30 am

    “we did not have any way to demonstrate if the mistakes the editor indicated really existed”
    or that the results reallzy existed.

  • Post a comment

    Threaded commenting powered by interconnect/it code.