The cell biologist Maria Pia Cosma, top-funded and prize-winning professor in Barcelona, is either an indestructible zombie scientist or a martyred saint. Her name means in Italian Mary the Pious, so it is probably the latter. So the following post is written in the style of a canonical satire.

Holy Mary Pia was accused by evil tongues on PubPeer of research misconduct in several of her publications. Philistines pointed crooked fingers and cast stones at duplicated western blot bands in papers published by St. Maria Pia in exalted scriptures like Cell. The accusers, envious of such impact factors, wished those gospels of scientific truth crucified on the cross of retraction and demanded for Our Lady of Barcelona to be judged and investigated. Yet Lord took mercy upon the tortured saint and sent His angels to defend her.

The martyr of research integrity, St. Maria Pia (right, standing) receives a sign for “Beatified”. Source: Cosma lab, under fair use

The first guardian angel the Lord sent was Kim Nasmyth, who used to hold his protective hand over St. Maria Pia when she served him as humble postdoc at IMP in Vienna. Nasmyth’s mighty name crowns three of most vilified papers by St. Mary Pia (Cosma et al 1999a, Cosma et al 1999b, Cosma et al 2011). This angel spoke to Retraction Watch:

“If someone comes to me with clear evidence for wrongdoing then I will be more than happy to respond. However, I am not aware that this has happened. For all I know, the allegations are part of a personal vendetta against Pia”.

Verily, for he who is without a sin shall cast the first stone. Then, the Lord sent a trinity of angels, Juan Valcarcel, Jaume Bertranpetit and Luis Serrano, the elders of the Centre for Genomic Regulation in Barcelona where St. Maria Pia leads her research-integrity-chaste laboratory. These elders brought forth a gospel of Josep Manel Rodríguez Sánchez, who cast his eye over a triptych of St. Mary Pia papers and exposed all those PubPeer-Philistines as godless liars who sinned in the face of the Lord by accusing this patron of research integrity of devious western blot bands duplications (see my earlier report).

The next angel sent by the Lord to defend the immaculate honour of St. Mary Pia, was Emilie Marcus, Editor-in-Chief of the holy truth journal of Cell and CEO of Cell Press. Angel Emilie declared that the absence of original data actually proves full reliability of St. Mary Pia’s research studies and announced to defend those papers with her editorial sword from any further blasphemous attacks, which lacked “the burden of proof” (see my other report here).

Now, Lord’s angels stayed the hand of another editor, Bernd Pulverer of EMBO Press. Two truthful gospels by St. Mary Pia were saved the fate of retraction and corrected instead (Zito et al EMBO J 2007 and Zito et al, EMBO Reports 2005). In the process, they were purged of all suspicions by removing tainted sections. These corrections were meant as partial retraction, a recent EMBO invention which allows authors to save their publications from full retraction by removing problematic figures as long as the central data remains unaffected.

For the paper Zito et al 2005, EMBO Reports issued on December 1st 2016 this authors’ note:

“We have come to realize that bands in Supplementary Fig 4A and B were derived from different gels and pasted on white background. Furthermore, several bands in Supplementary Fig 4A and B appear to be possible duplications (in particular, the bands in lanes 2, 6, 7 top panel of A and lane 2, bottom panel of B may have been duplicated; lane 7 bottom panel of A and lane 4, bottom panel of B may have been duplicated; lanes 5 and 6, bottom panel of A may have been duplicated), although this cannot be concluded unequivocally due to the low resolution of the figure. Supplementary Fig 4A and B displays loading controls of the co‐IP experiments shown in Fig 2A and C. These results, however, are not essential to support the main conclusion of Fig 2A and C, namely that SUMF1 and SUMF2 interact with themselves and with IDS and SGSH, because the reported interactions involved expression of tagged factors and, therefore, the interactions are unequivocally documented by anti‐tag‐specific antibodies. Furthermore, these interactions are confirmed by other results in the publication (mainly by the co‐IP experiment using differently tagged proteins in Supplementary Fig 3A, by the co‐IP of the endogenous proteins in Fig 2B, as well as by the experiments showing cysteine‐mediated interactions in Fig 3), through an unbiased mass spec method [1] and by others [2].

As source data are unfortunately no longer available and the content of these Supplementary Figures is not necessary for the conclusions of Fig 2A and B, we would like that Supplementary Fig 4A and B is considered withdrawn from the paper.

All authors concur with this statement and wish to apologize for the inconvenience caused”.

An Editors’ statement was added:

“In light of the potential image aberrations noted in the author’s statement, and since source data were not available for this figure, panels A and B of Supplementary Fig 4 are herewith retracted. These panels constituted loading controls for Fig 2A and C with limited impact on the qualitative conclusions made in the paper—see author statement for further details”.

Yet right away, the scheming PubPeer Philistines posted another sinful slander. The figure 3A, which is possibly just another trivial control for something else, was smeared with suspicions of duplicated bands and non-existent last lane:

The likely manipulated Figure 3A can probably be safely retracted as well, like most other figures of that divine masterpiece, without affecting its central findings. Yet the very figure 2A, which both authors and EMBO Reports declared as pristine and as cardinally paramount to this publication, was fingered for suspicion of band copy-pasting:

What will EMBO do now? Both St Mary Pia and her guardian angel Kim Nasmyth are EMBO members.

The correction for the EMBO Journal paper Zito et al, 2007 on December 1st 2016 sounded similar, with St. Maria Pia and coauthors admitting massive data manipulation and absence of original data:

“Figure 1D: We were made aware that the band marked as WT may have been spliced into this panel; however, this cannot be confirmed or disproved due to the absence of source data and the low resolution of the figure. This experiment confirms the results of experiments in Figs 1B and C, and 4C and D and F, and 5. Furthermore, the uptake of SUMF1 protein was also confirmed by mass spectrometry analysis (Supplementary Figure 1). This control was also intended to provide a reference for antibody recognition, which is also provided by the Western blot of protein extracts and media harvested from HepG2 cells (Fig 1D upper panels). Therefore, the possible splice does not affect the main conclusions of the paper.

Figure 4C (top panel): We were made aware of the overcontrasting of the Western blot result (4C; top panel). Considering other pieces of evidence in this paper (Fig 4A), our own additional results and the results from others (Dierks et al, 2003, Landgrebe et al, 2003, Preusser‐Kunze et al2005), where it was demonstrated that N141 is the only residue where SUMF1 is glycosylated, our findings have been validated in other contexts and the possible overcontrast of the Western blot data does not alter the conclusion of the paper.

Figure 4E: Fig 4E includes the analysis of Xenopus SUMF1 localization into the ER of HeLa cells, which appeared to be similar to the localization observed for SUMF1N141A (Fig 4B). We have come to realize that two areas of the immunofluorescence picture in the anti‐ERAB and merged panels of Fig 4E were blocked out. Although we cannot discern the data underlying the blocked areas because source data are not available, the correct localization of Xenopus SUMF1 in the ER was observed in other experiments (please see and confirmed by the observation that Xenopus SUMF1 was enzymatically active after expression in HeLa cells and it was correctly secreted but not taken up (Fig 4F–G). Overall, the results from the studies regarding Xenopus SUMF1 recapitulated what was observed for human SUMF1N141A (Fig 4B–D), and therefore, the conclusion of the paper stands. Nonetheless, we withdraw this figure panel, as it cannot be regarded as reliable in the absence of source data.

Figure 6B (upper panels): We were made aware of possible band insertions in the two right‐hand lanes of the SUMF1 panel as well as a possible duplication of the two right‐hand lanes of the beta‐tubulin panel. Since source data are not available, we cannot confirm or disprove these aberrations; however, the general conclusion of this experiment stands, because it is confirmed by the data presented in Fig 6A and B, bottom panels. Specifically, 6B includes experiments showing that the uptake of SUMF1 is mediated by the Mannose receptor in MEFs. This result is confirmed by the competition uptake experiments performed in human cells (Fig 6A) and in MPR−/−MEFs (6B bottom panels) where Mannan was used to inhibit the MR. We withdraw Fig 6B upper panels, as it cannot be regarded as reliable in the absence of source data.

Supplementary figure legend SI2A: While for most proteins, the samples corresponding to the medium and cellular pellets were run in the same gels; for SUMF2, the media and cell pellets were loaded in different gels, the Western blots were processed in parallel, and the two sets of lanes were collated together. This was not properly described in the Supplementary figure legend 2 which in the 5th line should read “The protein extracts and media were loaded in different gels and the filters were decorated with different antibodies: anti‐SUMF2, anti‐PDI, anti‐ERK and anti‐beta‐tubulin.”

All authors concur with this statement and wish to apologize for the inconvenience caused“.

The EMBO J editors decided:

“In light of the potential image aberrations noted in the author’s statement, and since source data were not available for the figure panels in question, Figs 4E and 6B, upper panels, are herewith withdrawn”.

Yet the PubPeer Philistines were quick with new slander of duplicated western blot bands against St. Mary Pia:

Hail Mary Pia full of Grace, the Lord is with thee. Blessed are thou among women and blessed is the fruit of thy mind published in Cell and other elite journals. Amen. 

Update 17.12.2016. The two EMBO Corrigenda blessed also the first-author disciple of St. Maria Pia, Ester Zito, who is now group leader at Istituto di ricerche farmacologiche Mario Negri in Milan. This young and utterly innocent scientist was granted by the Italian charity foundation Telethon over half a million Euro to help find cure for degenerative muscle disease. Thanks to the glorious kindness of EMBO, the charitable private donors of economic crisis-struck Italy will not have to vex if they might have given their small savings to a ruthless cheater. Amen, again.

6 thoughts on “Two EMBO corrections for the martyred saint Maria Pia Cosma

  1. [The user Plantarum had been dis-invited from commenting on my site, but this comment is too funny to be deleted. -LS]
    “the following post is written in the style of a canonical satire”
    Confirming thus that Leonid Schneider is the de facto anti-Christ.


    1. Dear Vicki, thanks a lot for commenting. In fact, EMBO introduced the model of partial retraction for… tada: Olivier Voinnet.
      The editorial by Bernd Pulverer appeared in the same EMBO J Volume 34, Number 20, 14 October 2015 issue as these 3 Corrigenda:
      Two classes of short interfering RNA in RNA silencing
      Andrew Hamilton, Olivier Voinnet, Louise Chappell, David Baulcombe
      The corresponding author was alerted through the PubPeer website to background pixel pattern duplications in Figure 4B of this paper. The Figure shows that a viral suppressor of silencing (P25) blocks 25‐nt siRNA and systemic silencing.

      Nuclear import of CaMV P6 is required for infection and suppression of the RNA silencing factor DRB4
      Gabrielle Haas, Jacinthe Azevedo, Guillaume Moissiard, Angèle Geldreich, Christophe Himber, Marina Bureau, Toshiyuki Fukuhara, Mario Keller, Olivier Voinnet
      Figure 4D, lower panel: the images corresponding to the P6m3 and P6ΔdsR mutants are erroneous duplications of images (or sections thereof) in Figure 4A and 5A. Co‐author Olivier Voinnet made these errors during figure mounting. Since the source data were unfortunately unavailable, the experiment depicted in Figure 4A–D was reproduced by retrieving seed stock of the corresponding transgenic plants and growing them side‐by‐side under identical, controlled conditions (Geldreich A, Himber C, Haas G, Keller M and Voinnet O (2015) Phenotypic analysis of Arabidopsis transgenic plants constitutively expressing the P6 protein from Cauliflower mosaic virus or mutant alleles thereof. bioRxiv doi: 10.1101/023283). These results concur with the original conclusions.
      Figure 5A, lower panel: the image of the P6ΔdsR mutant is an erroneous duplication of the second image in the upper panel. Co‐author Olivier Voinnet made this error during figure mounting. The source files have been retrieved to produce a corrected Figure 5A whose conclusions remain unchanged.
      The properly assembled Figure 5A is displayed below alongside the source data.
      Figure 5A.

      Pictures used in Figure 5A.

      Figure 5F (DRB4‐P6 co‐immunoprecipitation): because the panels are excerpts from the same gel, adequate dividers should have been added during mounting of panel F by co‐author Jacinthe Azevedo. The source file has been retrieved to produce a corrected Figure 5F; the conclusions remain unchanged.
      Figure 5F.

      Original data and assembly of Figure 5F.

      Note that the partial reproduction of images Figure 3C and D, left panel was intended but was not specified in the legend of Figure 3.
      All authors concur with this statement and wish to apologize for the inconvenience caused.
      Editors’ statement
      Figure 4D contained duplicated images derived from Figures 4A and 5A. Figures 4A and 5A have been verified, but since source data were not available for Figure 4D, this specific panel is herewith retracted. See author statement for further details.
      We alert readers to the fact that a number of related papers are also subject to a corrigendum or to a retraction.

      Differential effects of viral silencing suppressors on siRNA and miRNA loading support the existence of two distinct cellular pools of ARGONAUTE1
      Gregory Schott, Arturo Mari‐Ordonez, Christophe Himber, Abdelmalek Alioua, Olivier Voinnet, Patrice Dunoyer
      Figures 3, S1 and S2A: The rRNA loading control depicted for these three figures (originating from the same membrane probed and stripped multiple times) is not the cognate one: it corresponds to a loading control mistakenly reused from a previous publication of our group (Nat Genet 39: 848–856, 2007). Source data for these figure panels are displayed below.
      Figure S2B: A divider should have been added to clearly indicate splicing of the original blot made to remove one of the two HC samples present on this blot. Source data relevant to this figure are displayed below. Note that source data for the @822 probe were not available, but a source data image of the same membrane hybridized to the @171 probe is displayed alongside a replicate experiment with the @822 probe.
      Figure S3: Dividers should have been added to clearly indicate splicing of the original blots. In addition, incorrect splicing of the total RNA blots for @255 and @TAS3 led to a duplicated band mistakenly used for the final figure. Source data for this figure are displayed below.
      Figure 5A (@AGO1 blot): Background duplication was inserted during image editing to cover handwriting present on the original blot. Source data for this figure are displayed below.
      Figure 5B (@CHS blot): Background duplication was inserted during image editing to cover handwriting present on the original blot. In addition, the rRNA loading control is not the cognate one and includes two sets of duplicated bands. Source data for this figure are displayed below.
      Source data for Figures 3, S1 and S2A.

      Source data for Figure S2B.

      Source data for Figure S3.

      Source data for Figure 5A.

      Source data for Figure 5B.

      The conclusions drawn from these figures and from the paper remain unchanged. Patrice Dunoyer, who assembled the figures, takes the responsibility for these errors. All authors agree with this statement and wish to apologize for any inconvenience caused.
      Editors’ statement
      The authors provided to the journal facsimile images of relevant labbook pages and source data to explain the apparent manipulations; these data support the conclusions made, as described by the authors. For Figure S2B, source data of the same experiment (hybridized to the @822 probe) were not available, but a source data image of the identical membrane hybridized to the @171 probe and a replicate of the experiment as well as total RNA controls were made available.


  2. Is it bad of me to wonder, if the paper’s conclusions remain unaffected by the removal of so much of the supporting data, why were those data included in the paper in the first place? As padding? Merely corroborative detail, intended to give artistic verisimilitude to an otherwise bald and unconvincing narrative?

    Liked by 1 person

    1. The 3rd paper CRG “investigated”, was this one, Cosma et al Mol Cell Biology 1998.:
      That one was not corrected or even commented upon by the journal.

      To me it now seems that any attempt to expose any band duplications in Cosma’s papers would be a direct affront to one of the most prestigious research institutes in Europe, the CRG. Thus, by officially sanctifying Cosma as “Our Lady of Barcelona” and her papers as immaculately conceived, CRG created a religious dogma despite all objective facts, which the faithful research community must follow or face excommunication. Amen.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s