This text was first published on September 30th as Spanish translation on Hipertextual.
The Spanish Institut d’Investigació Biomèdica de Bellvitge (IDIBELL) in Barcelona has discovered a new application of the famous Schrödinger uncertainty theory, by extending quantum mechanics from single atoms to entire scientific publications and its authors. The traditional Schrödinger cat inside a box with a poison-releasing radioactivity detector can be simultaneously dead and alive, and you never know which until you open the box. According to IDIBELL, a senior scientific research group leader can be also simultaneously 100% responsible and utterly not responsible for any given scientific publication of his, depending on the nature of observation. Such a “Schrödinger cat”-scientist is Manel Esteller, cancer researcher and director of strategic projects at IDIBELL. On the one side Esteller is being hailed and awarded with highest prizes for his publications on cancer genetics (most recently the Gold Medal of Honour by the Catalan Parliament), from the other side he is being declared as entirely not responsible for these same research papers, as soon as any evidence for data manipulations in them surfaces. Esteller’s quantum state of responsibility for his own research oscillates between yes and no depending on whether you open his IDIBELL box in order to give him a prize or to report suspected research misconduct.
This is how the Barcelona Theory of Quantum Irresponsibility developed.
The IDIBELL cancer biologist Esteller used to be the former PhD advisor of the by now infamous Portuguese scientist Sonia Melo. Several of Melo’s research publications on the molecular mechanisms of cancer were flagged for suspected data manipulations on the website PubPeer in 2015, most cases were about duplicated images and datasets, occurring in a manner which strongly suggests a manipulative intent to deceive. These suspicions led the European Molecular Biology Organisation (EMBO) to re-examine the prestigious Young Investigator award which Melo was granted in December 2015. Already on February 29th 2016, EMBO announced to withdraw Melo’s funding after having determined significant problems with her publications. The Portuguese scientist was suspended as leader of her own research group and is currently investigated for suspected research misconduct by her new Portuguese employer, Instituto de Investigação e Inovação em Saúde (I3S) at Universidade do Porto.
Three of these problematic Melo papers came from the Manel Esteller lab, one was in fact even retracted from the prestigious journal Nature Genetics on January 27th 2016, due to image duplications. No direct blame was apportioned to any of the authors in the retraction notice. As the last and the only corresponding author, as well as doctoral thesis supervisor, Esteller would normally be fully and entirely accountable for this publication. However, the IDIBELL group leader was quick to distance himself from his own Melo et al Nature Genetics 2009 paper just before it was felled by the retraction. He placed on November 25th 2015 a comment on PubPeer, where he shifted the entire responsibility on his first author, Melo. At that time, it is highly probably that the decision to retract the paper was already made by the journal. In this vein it is quite understandable why Esteller made no such comments about his and Melo’s other two problematic publications in the elite journals Cancer Cell and PNAS, and indeed neither has been retracted or even corrected yet. There, the IDIBELL senior researcher kept silent and preferred to leave the credit for the impactful publishing achievement to himself.
Neither did he comment on any of the further three publication from his own lab, where Melo was not even one of the co-authors. The accusations ranged there from concealed inappropriate splicing of DNA and protein gels to duplication of gel bands, possibly suggestive of intentional falsification. In one case Esteller had to issue in July 2016 a correction admitting gel lane splicing, but neglecting to discuss their suspiciously uneven nature.
What Esteller did instead of facing his responsibilities, was to remove his own suspicions-tainted research from his official list of publications at IDIBELL website. Not just the retracted Melo et al Nature Genetics paper was purged from his lab’s wall of fame: also the other two Melo papers in Cancer Cell and PNAS were delisted in spring 2016 (evidence here). Apparently, Manel Esteller’s research group should never again be associated with Sonia Melo or in fact research misconduct.
The case seems also closed for IDIBELL. In February 2016, the institute’s director Jaume Reventós made clear to me that for all Esteller publications featuring his PhD student Melo only the latter was solely responsible:
“As you know, since many years, Sonia Melo is not any longer in our institution and she is pursuing her career elsewhere. Concerning the publications in which Dr. Melo is not an author, I want to let you know that we will pursue our inquiry following our standard procedures for those matters”.
Soon after Hipertextual was informed by IDIBELL what kind of inquiry it will be: namely with the lab head tasked with investigating himself. There will be no other investigators from IDIBELL or from outside invited to the sessions of tough questioning which Esteller was supposed to submit himself to in front of a mirror. The Schrödinger quantum uncertainty box was opened and the corresponding author cat inside was declared to be utterly not responsible for the data integrity in any of the papers from his own lab:
“The group of Dr. Esteller is reviewing all results which could be subject to dispute, as it is common in scientific practice, in response to requests from independent researchers requiring this information.This process is completely normal and is made group-internally according to the usual protocols of publication and scientific discussion […].
It is important to emphasize that the responsibility for managing the results of a scientific publication usually falls on the first author, and that the work of the senior author (last name), is to direct, coordinate and review the work, which occasionally allows the published errors to pass unnoticed through his filter. Thus, the attitude of Manel Esteller has been exemplary in recognizing a possible error and retracting the work quickly, once doubts were expressed about it”.
However, the quantum uncertainty works both ways. Whenever the box was opened with an award or a prize in mind, the cat inside suddenly clawed all the credit for his publications as corresponding author. Indeed, Esteller received many awards before and even after the Melo scandal broke, all of these awards served to recognise his exceptional research output published in prestigious journals… such as the ones which are now strongly suspected on PubPeer of deliberate data manipulations.
Just when the retraction of his paper with Melo was prepared in December, Esteller received the National Research Prize 2015 from the Government of Catalonia and the Catalan Foundation for Research and Innovation, for his research in epigenetics of cancer. In July 2016, the IDIBELL scientist received the 28th Catalunya International Prize “for his revolutionary task, especially on oncology research, with what they significantly contributed to the advances of global medicine”. In the same month, he was showered with funding from the European Union and its Research Council to study epigenetics of cancer. Most recently, in September 2016, Esteller was awarded the Gold Medal of Honour by the Catalonian Parliament, for his “fight against cancer”. Not a single junior contributor from Esteller’s lab was ever mentioned in this ceremonial context, much unlike when the data integrity concerns in the same papers was discussed. Instead, all “revolutionary” results achieved by the first authors of his problematic publications on this very field of cancer gene regulation suddenly became exclusively the lab head’s own discoveries and responsibilities.
The Catalonian cat inside the box is simultaneously both responsible and not responsible for his own papers. However, this conundrum is far from singular, in fact another former mentor of Sonia Melo found himself inside exactly the same quantum uncertainty. After completing her PhD thesis with Esteller, the young Portuguese researcher went for postdoctoral training to the lab of Raghu Kalluri, at the prestigious MD Anderson Cancer Center in Texas. Several of Kalluri’s publications are flagged on PubPeer for suspicions of data irregularities; five of them feature Melo as first or second author. However, also Kalluri is not being investigated by his research institute, as I found out. And this is despite (or because) of the fact that one of his highly problematic papers, namely Melo et al Nature 2015, was seminal in raising the staggering amount of at least $80 Million for commercialising a novel cancer diagnostics method which he and Melo have discovered. Kalluri’s own private company and his employer MD Anderson are directly involved into the marketing of a cancer biomarker technology which many decried on as irreproducible or even as possibly dishonest.
As it turned out, IDIBELL is in a very similar situation of financial conflict of interests: the Catalonian institute runs also some heavy business activities with Esteller, which might explain the oscillations of the quantum uncertainty in regard to the responsibilities for his own research. Together with the also Barcelona-based pharma company Ferrer, Esteller has developed a cancer diagnostics test called EPICUP (see press release and The Lancet Oncology paper Moran et al 2016). A patent was filed in 2012 and approved in January 2016, the inventor is Esteller and the applicant is IDIBELL. The first author of the relevant publication suddenly seems utterly irrelevant and is not mentioned anywhere in the patent or even the press release. Yet this might change, should questions arise about the reliability of this cancer diagnostics test. What the cat inside the Catalan Schrödinger box is really up to, depends on the observer and what this observer is offering.
Update 2.11.2016. Nature Communications issued today a correction for an Esteller paper (Lopez-Serra et al 2014) where the authors admitted inappropriate data presentation and loss of original data. Sonia Melo was not a co-author of this paper.
Update 22.01.2017. IDIBELL now published on PubPeer the conclusions of their internal investigation into data irregularities in Esteller lab papers, which was led by Ombudsman Francesc Xavier Bosch. In a 180° change from the previous standpoint by IDIBELL, Esteller became now responsible for his lab’s publications after all. But this possibly only because no misconduct was found, only technical issues. A quote from the full statement:
“After reviewing the information provided, the panel of experts forming the commission has concluded that there is no sign of scientific fraud or data manipulation besides technical negligence in figure assembling by the authors. Although Dr. Esteller, as corresponding author and supervisor, has the final responsibility for the quality of the results published, no deliberated intention of fabricating results was observed so the case presented cannot be related in any form with research misconduct”.
As I previously commented the $80 Million COI Kalluri/Melo paper it is not reproducible:
More specifically, the conclusions of my brief analysis of my mass spectrometry data concerning the manuscripts:
Melo et al. 2015. Glypican-1 identifies cancer exosomes and detects early pancreatic cancer. Nature. 2015 Jul 9;523(7559):177-82
Costa-Silva et al. 2015. Pancreatic cancer exosomes initiate pre-metastatic niche formation in the liver. Nat Cell Biol. 2015 Jun;17(6):816-26
I cannot find GPC1 in exosomes derived from pancreatic cancer cell lines such as BXPC3 and MiaPaca2, neither I found GPC1 in plasma derived from pancreatic tumor-bearing mouse or control mouse. Indeed, is very difficult to find GPC1 in any exosome type.
Indeed, MIF is present in the exosomes of several human pancreatic cancer cell lines but not in the exosomes derived from PANC1 cells. MIF is also present in the exosomes derived from normal mouse pancreas exosomes and also I cannot found MIF in plasma derived from pancreatic tumor-bearing mouse or control mouse. Finally, MIF is not highly expressed in PDAC exosomes (MIF has a similar expression in normal mouse pancreas exosomes)
Awesome article Lenoid.
It looks like that Esteller have obtained almost a million euros in cash (in pocket) money (in multitude of high profile prizes in Spain), this sum add to his high salary. So, to do this kind of manipulative research is economically viable for him, because if there any problems he just was cheated by the first-author student.
By the way, the paper of Melo is just one of the many papers that Esteller has in pubpeer.
You are doing an amazing job, Leonid. Honest science need more journalists like you. And harder policies on those faking studies, or evading responsibilities like mr. Esteller did. To get grants we have to compete against overinflated CVs [edited, -LS] after all…
Another prize money for Esteller.
He gets the money, his PhD students get fired… It looks a good deal.
Pingback: Sonia Melo fully exonerated and reinstalled as PI by her Portuguese employer I3S – For Better Science
Update 22.01.2017: is this surprising?
Pingback: Melo and Kalluri defend discredited Nature paper with preprint, where they admit data “adjustments” – For Better Science
Pingback: Nature rewards data manipulation with a Mentoring Award – For Better Science
Pingback: Zombie scientist Sonia Melo awarded by AstraZeneca – For Better Science
Pingback: The Perennial Northern Blot of Lopez-Otin – For Better Science
Pingback: David Latchman, the reckless Master of Birkbeck – For Better Science
Pingback: How Andrea Cerutti molested and defiled Journal of Immunology – For Better Science
Pingback: The impossible Toni Camins – For Better Science