A former Vanderbilt University biomedical engineer committed fraud on a massive scale, according to a new Office of Research Integrity (ORI) report.
Igor Dzhura is banned from receiving federal funding for three years, and is retracting six papers, which have been cited more than 500 times. Since leaving Vanderbilt, he has worked at SUNY Upstate Medical University, and now works at Novartis.
According to the ORI, Dzhura was a busy boy at Vanderbilt, faking images and drastically inflating the number of experiments he conducted by duplicating computer files and saving them in nested folders. The total body count from his work includes:
…submitting and publishing multiple falsified and/or fabricated action potential traces and summary data in at least sixty-nine (69) images in twelve (12) different figures across seven (7) publications and three (3) grant applications by duplication and relabeling of traces; resizing, modifying, and splicing different traces; and modifying and/or duplicating bar graphs.
Dzhura admitted to the ORI that he had
…significantly departed from accepted research practices by engaging in the intentional and knowing fabrication and falsification of data files.
Here’s a taste of what he did, according to the ORI:
- created a hierarchy of computer folders containing duplicated and renamed files; the falsified groups of files included eighty-two (82) groups of duplicated files with each group containing two to twenty-one (2-21) duplicates, which made it appear that experiments were conducted when they were not; and
- used the falsified and/or fabricated data files in Figure 6 of a paper published in the American Journal of Physiology-Heart and Circulatory Physiology (292(5):H2202-H2211, 2007), to represent Ca+ currents in cardiac myocytes from CLCAD-/- mice; specifically, Respondent claimed that Figure 6 represented results from seven (7) mice when the data files were three (3) sets of duplicated and renamed files plus one additional data file. All of the data files were part of larger groups of identical duplicated and renamed data files on the Respondent’s hard drive.
He also agreed to six retractions. We’ve added the number of citations next to each link, according to Thomson Scientific’s Web of Knowledge:
- Calmodulin kinase determines calcium-dependent facilitation of L-type calcium channels, Nature Cell Biology, 247 citations
- Calmodulin kinase and a calmodulin-binding ‘IQ’ domain facilitate L-type Ca2+ current in rabbit ventricular myocytes by a common mechanism, Journal of Physiology, 2000, 46 citations
- Calmodulin kinase II and arrhythmias in a mouse model of cardiac hypertrophy, Circulation, 2002, 159 citations
- Cytoskeletal disrupting agents prevent calmodulin kinase, iq domain and voltage-dependent facilitation of l-type ca2+ Channels, Journal of Physiology, 2002, 40 citations
- C terminus L-type Ca2+ channel calmodulin-binding domains are ‘auto-agonist’ ligands in rabbit ventricular myocytes, Journal of Physiology, 2003, 16 citations
- A dynamic α-β inter-subunit agonist signaling complex is a novel feedback mechanism for regulating L-type Ca2+ channel opening, FASEB Journal, 2005.
After Vanderbilt, Dzhura moved on to SUNY Upstate Medical Center. We emailed his principal investigator there, George Holz, who told us:
Dr. Dzhura left my lab back in 2010* and has been working at Novartis in Boston although I have had very little contact with him. I was unaware that ORI was conducting this investigation. Thank you for bringing it to my attention. When Dr. Dzhura was working at my lab in 2009 and 2010, there was no mention from him or any of his previous employers that his prior work had been questioned.
*Update 6:30 p.m. Eastern, 11/20/14: We’ve changed a date in Holz’s statement to reflect that Dzhura left the lab in 2010, not 2011 as originally stated.
Update 1:20 p.m. Eastern, 11/21/14: A Novartis representative reached out to inform us that they’ve fired Dzhura after discovering he included the faked papers on his application:
We have learned that Igor Dzhura included papers with fraudulent data in his application for employment at Novartis. Falsifying data is not acceptable and we have terminated his employment with the company. We are conducting an internal review to ensure that there was not any scientific misconduct related to his research here.
“…banned from receiving federal funding for three years…”
Is that all? For admitted deliberate falsification why isn’t he banned for life? For committing fraud why isn’t he fined and/or imprisoned? Imo, the timid “meh, so what” sanction does nothing to protect the public, and is every bit as shameful as the behavior itself.
That was my very first thought. He should be banned for life, and he should be fired.
“significantly departed from accepted research practices”
2014 speak for a fraudster, as in “I significantly departed from accepted bank-client relations when I help up the bank at gunpoint.”
Does this work as a defense in a law court 🙂
Does the funder get their cash back,so that people who actually do experiments can progress their work?
ORI has no prosecutorial authority. Blatant cases such as this one would (should) be referred to the Attorney General for prosecution. however, it appears as though the AG only prosecutes when a Senator writes a well publicized letter.( Grassley. U. Iiowa. ’13 ?)
Trainees cheat because they are fearful that they won’t get an academic appointment. A good mentor might mitigate that.
Faculty seek fame and fortune (grants, promotion)). Were they more concerned that their act of misconduct could be prosecuted as Federal crimes punishable by imprisonment and/or a hefty fine, they might reconsider fabricating or falsifying
data. Perhaps if the scientific community itself asked the appropriate Federal authorities to invoke more severe penalties on the rogue scientist, they might take such fraudulent behavior more seriously.
Don Kornfeld
Right, Donald, ORI has no prosecution authority. The United States Department of Justice is the federal prosecutor that would review serious criminal and civil misbehavior for possible prosecution.
To N.E.W.’s comment, the U.S. Government standard debarment period is 3 years – however, for very serious cases of research misconduct in HHS research, ORI/HHS has imposed or gotten settlement agreements to impose 5 to 10 years debarments, as well as lifetime debarments — see my earlier RW posts with examples of them at:
http://retractionwatch.com/2014/04/25/former-mount-sinai-postdoc-faked-gene-therapy-data-ori/#comment-95453
http://retractionwatch.com/2013/12/06/should-scientific-misconduct-be-handled-by-the-police-its-fraud-week-at-nature-and-nature-medicine/#comment-69839
Kudos to Dr. Kornfeld. This type of situation has been going on far too long as faculty seek “…fame and fortune…” The punishment does not seem to fit the crime, and as long as there are no major penalties, this will continue. A related question is what have the institutions, whether universities or companies, done in such cases? To the best of my knowledge, nothing much. They need to take this situation much more seriously, and think about the environment that leads to or even feeds such behaviors.
Significantly higher numbers of citations are reported by Google Scholar compared to Web of Knowledge. For the most cited (now retracted) paper, WoK gives 247 citations while GS reports 329 citations.
Relying on GS, the pack of retracted papers accumulated more than 700 citations.
Two good recent reminders of why we rely on Thomson Scientific rather than Google Scholar:
http://retractionwatch.com/2014/11/17/fake-citations-plague-some-google-scholar-profiles/
http://scholarlyoa.com/2014/11/04/google-scholar-is-filled-with-junk-science/
Though this is true, its also worth keeping in mind that Web of Knowledge does not index papers until their “official” publication date. So, not only are “in press” papers not included, but even if there is a paper that has been assigned to, say, the February issue of a journal and is currently available in its final form with page numbers assigned, in my experience, WoK won’t show it until February, even if the entire issue is complete and the journal is now working on March. Issue dates are rapidly becoming a fiction, at least in my field; if there is fault here, it belongs to the journals and not WoK, but the result is that you can’t rely on WoK to keep up with the latest papers–by the time they hit it, they are likely to be several months, if not a year or more, stale. (The papers in this case may be old enough that there aren’t many legit citations in the pipeline, but it could explain part of the difference.)
@Duvane
You’re certainly right, the army of Google’s robots should be more responsive than WoK regarding virtually published papers.
I also realize that, apparently, Google Scholar never flags retracted papers. Once a paper has been published, the link will survive in GS ad infinitum, the paper remains accountable for h-index computations, etc. And, more worrying, the paper will be cited, even after retraction.
I’ve seen the following case today: a communication in “Crystal Growth & Design”, edited by the Royal Society of Chemistry:
Published online: 22/02/2006
Retracted with an uninformative one-line notice: 13/04/2006
Until now: 19 citations in GS, 9 in WoK. The paper remains in the institutional CV of the main author, and is available from his university e-print repository (behind a login). The ESI and crystal data are available, and the pdf is available-for-sale on the RSC site (£36), and downloadable for subscribers.
It’s actually difficult to notice that the paper has been retracted 8 years ago: the only evidence is a small link in the midst of the advertisements, which reads “correction”.
What are Novartis going to do about this? Will they continue to employ a known fraudster?
The fudged KM-curves of the mice with brain tumors probably keeps them already occupied.
http://retractionwatch.com/2014/11/17/retraction-appears-for-faked-study-of-novartis-anti-cancer-compound/
Dzhura’s “supervisor” at Vanderbilt was Mark Anderson who went onto be chair of medicine at Iowa and is now chair of medicine at Johns Hopkins.
Wow. That makes Prof Anderson a serious player in US biomedical research. Has he said anything about Dzhura’s work in his lab?
Not that I am aware of. But essentially all of the high profile work from his lab while he was at Vanderbilt is implicated in this misconduct finding. And presumably it was this record of work that, at least in part, contributed to him becoming division chief of Cardiology at Iowa and then Chair of Internal Medicine at Iowa before he moved to JHU.
Once again it appears as though a lab minion is solely responsible when things go bad while the lab head’s career forges ahead on the basis of questionable research.
The PI shares the credit and doesn’t share the blame – no surprise there…
As you can see, he wasn’t shy about taking credit for the retracted work:
http://www.mc.vanderbilt.edu/reporter/index.html?ID=1079
Safe firing in legal terms I would guess, because his CV was falsified. I wish universities would be as robust as Novartis – instead universities seem to “part by mutual agreement”, as happened at Liverpool when Melendez left (note Melendez had not published any falsified data during his short time at Liverpool).
When I read that Novartis had hired him, I wondered if they would have the grounds to fire him. Falsifying an application definitely fits this (pharma companies actually have strong ethics requirements, albeit there are employees that violate them).
It’s very heartening to hear that Dzhura isn’t profiting from his fabrications. Thumbs up for Novartis.
Like I said, Dzhura’s “profit” from his fabrications is relatively modest in comparison to the extensive profits reaped by his “mentor” Dr Anderson.
There has also been a recent “brief communication” in Nature contesting work subsequently published by the Anderson group that was, at least in part, based on or prompted by the Dzhura publications that are now in question.
http://www.nature.com/nature/journal/v513/n7519/full/nature13626.html
I think there could be more to this case than just a rogue post doc making things up all by himself.
I agree. Always the small fish are caught. The big fish have support and protection. Such incidents occur on a daily basis everywhere out of pressure or others in the lab volunteering to produce something that will work in their favor with the mentor. The ban should be on the lab director. The firing should be of the lab director. The lab directors are so full of themselves and not interested in the everyday workings of the lab – they are the most cruel people asking for data all day long to further their own cause rather than further science. There are no true scientists left in this world.
Not surprisingly, Anderson will deny all knowledge of the fraud and distance himself from Dzhura. But when Dzhura was his star performer at Vanderbilt he was happy to take the credit. And a large part of Anderson’s meteoric career was built on work coming out of his lab at Vanderbilt. And a lot the federal grants that Anderson got were based on Dzhura’s data. Notably, Jeffrey Balser was also co-author on the retracted work. Balser is now CEO of Vanderbilt Health making millions. Wall Street all over again-the big fish hardly ever are held accountable.
Certainly would be interesting to know if JHU were told about the Dzhura investigation during Anderson’s recruitment and/or if this had anything to do with his motivation for leaving Iowa? Also, the Vanderbilt inquiry was resolved in part by an admission of guilt from Dzhura which is somewhat unusual in these kinds of cases but clearly represents a convenient and tidy outcome for Anderson and Balser.
Reading this disturbing material makes me wonder how much more of this sort of fraudulent activity is taking place in connection with other medical labs across our nation that nobody knows about et.
Another one – Expression of Concern. Note also the co-corresponding/co-senior author on this one, who was recruited by Anderson, and who supervised most of this work and put paper together – Anderson was already at John Hopkins when this paper was published.
https://pubpeer.com/publications/E5C8180E256D9238251ABBFC4B4EE4