A nearly ten-year-long series of investigations into a pair of plant physiologists who received millions in funding from the U.S. National Science Foundation has resulted in debarments of less than two years for each of the researchers.
The NSF Office of Inspector General recently posted its close-out report on its decision and a review of the University’s investigation, which had recommended a total of eight retractions or corrections. Although the investigator’s names have been redacted, the text of retractions and corrections quoted in the report corresponds to papers by Jorge Vivanco and his then-postdoc Harsh Bais at Colorado State University.
Bais has since joined the faculty of the University of Delaware, where he is now an associate professor, and obtained $2.5 million worth of grants from the NSF and over $500,000 from the National Institutes of Health. Vivanco has been promoted to a full professor, and received more than $3 million in NSF funding along with awards from the U.S. Department of Agriculture.
We first wrote about the researchers in 2011 when they retracted their 2005 Nature paper after “a key reference” (also co-authored by the pair) was retracted after they were “unable to find experimental data.” This note, plus the appearance of some other retractions, made us wonder if there was more to the story. The OIG’s answer: Most definitely.
This story, like many others, begins with a finding that was too good to be true. In the April 2002 issue of Plant Physiology, Bais and Vivanco promised to “unravel part of the mystery” of why spotted knapweed is such a nasty invasive plant. According to the authors, the plant’s roots secreted a noxious compound called (-)catechin that caused plants around it to wilt and die. A 2003 New York Times article about their follow-up paper in Science was headlined “Forensic botanists find the lethal weapon of a killer weed.” In a report to their USDA funders, the team summarized the impact of their findings:
Our studies have been highlighted in popular newspapers, magazines, and TV news shows, including The New York Times, Scientific American Magazine, National Geographic, CNN News, and the Discovery Channel. As a result of our studies a company has licensed one of the allelochemical compounds as an ecologically-benign herbicide.
But by October 2004, the duo’s collaborators within CSU, and at the USDA research center, began to have their doubts about research on this root compound and others, according to the OIG report. The batches that Bais was preparing were “inconsistent in color and smell” and contained “extraneous compounds.” Although the University terminated a commercial licensing award, Vivanco “continued to request and to obtain federal funding from NSF and other agencies . . . while suspecting or knowing [the research] to be irreproducible.”
CSU faculty members raised allegations of misconduct to the Dean of the College of Agricultural Sciences, and the Dean asked Vivanco’s department chair to conduct “several interviews and report back to him.” As the NSF report puts it,
An informal, undocumented evaluation of the allegations at the College level appeared to evolve into an informal inquiry outside of University 1 ‘s prescribed policy, including analyses of the batches of the Mixture and blind tests of the extraction methodologies reported in Article 1.
During that preliminary investigation, the department chair (Witness 13) recommended to Vivanco that the issue “be worked out in the literature,” according to the report. (The department chair has not yet responded to an email and phone call requesting comment.) In the spring of 2006, the report says Vivanco got much the same answer when he initiated a meeting with the College Research Integrity Officer to discuss the lingering problems with Bais’s data. The College RIO’s “general recommendation to [Vivanco] was that nothing could be proven.”
Ultimately, rather than retracting the papers with the suspect data, the OIG says Vivanco tried to “save-face” and downplay the earlier results in subsequent publications even as he tried to advance the story in grant proposals and publications. In 2007, he coauthored a paper that failed to replicate Bais’s findings in the field:
Although previous consistent reports of high soil catechin concentrations provided circumstantial evidence for a role of catechin in C. maculosa invasion (Bais et al. 2002, 2003; Perry et al. 2005b; Thelen et al. 2005; Weir et al. 2006), this more extensive study, together with those of Blair et al. (2005, 2006), suggests that high catechin concentrations rarely occur in C. maculosa soil. More confidence can be placed in the current results. Thus, the infrequency of soil catechin weakens the hypothesis that it plays a role in C. maculosa invasions.
The paper pointed that Bais’s work had not used controls and suggested that could be the source of the discrepancy. All of this was enough, evidently, to pique the interest of the NSF investigators, who began interviewing current and former CSU faculty and staff and learned about the preliminary investigation. The OIG requested that CSU look into the misconduct allegations once again, and, it’s at this point that CSU appointed an investigation committee, as required under its research misconduct policy.
This formal investigation concluded that Bais had intentionally falsified and fabricated data, while Vivanco had been complicit in research misconduct by capitalizing on the results after learning they were false. The CSU committee recommended that Vivanco be “considered for demotion by one academic rank” because he knew that the science used in his promotion was “highly suspect.” CSU, however, declined to demote him after a review. In 2014, NSF debarred Vivanco for three years and Bais for five, but, following an appeal, it has since set their terms to expire in May and June 2016, respectively, according to records in the System for Award Management.
All told, we’ve identified a total of
five six retractions, along with a string of corrections. Excerpts from notices in the OIG report match language present in the pair’s notices in Science (correction), Phytochemistry (retraction), and the Journal of Agricultural and Food Chemistry (withdrawal). In addition to the Nature paper, the duo have also retracted a second paper in Phytochemistry, along with a paper in Plant Physiology, and they published corrections to papers in Journal of Chemical Ecology, the Proceedings of the XII International Symposium on the Biological Control of Weeds and the Journal of Agricultural and Food Chemistry.
Bais could not be reached by telephone and failed to respond to an email from Retraction Watch, but in his response to the investigation, which is included in the OIG report, he criticized it for relying on “circumstantial evidence.” Over the telephone, Vivanco said he was unaware that the report had been posted online, but declined to comment other than saying that the retractions are “a matter of fact.” He later provided a statement to Retraction Watch over email:
I cannot comment on the specifics of this as it involves personnel matters for researchers who may or may not still be at CSU. Personally, I cooperated fully in a process that I think worked to correct the scientific literature and am confident that all publications for which I am the corresponding author are accurate and reflect the highest standards of research integrity. While there are aspects of the outcome that I may not necessarily like or agree with, I have learned some important lessons which have made me a better PI going forward.
In a statement to Retraction Watch, Alan Rudolph, the Vice President for Research at Colorado State University had this to add:
It is CSU’s stated policy that we cannot speak to specifics with respect to allegations of research misconduct. I can say that everyone involved in the adjudication of any allegation of misconduct takes their role very seriously and we have a strong commitment to research integrity. This includes proactive training and mentoring of our students, post docs and junior faculty. Our faculty are committed to maintaining the integrity of scientific inquiry and the scientific record.
Independently, Vivanco also retracted a 2012 paper in Frontiers in Plant Science, which he attributed to “mislabeling and duplication of images.” That paper did not include Bais as a coauthor.
Update 9/3/15 7:56 a.m. eastern: Bais sent us a document that he claims provides support for his research, which you can view in full here. He also sent us a statement, in which he says he’s appealing the NSF’s decision:
While I strongly disagree with many portions of the NSF’s Final Determination, it is important to note that it unequivocally states that my conduct was unintentional: “ . . . your misconduct was not intentional and you did not have primary authority or the CSU laboratory where the events occurred”.
My attorney, Barry Nelson Covert, Esq., and I have filed a request for reconsideration with the NSF and our legal team has filed an appeal of the NSF’s Final Determination with the United States District Court for the Eastern District of Virginia. We content that absent a finding of intent or willfulness, the NSF simply cannot find that there was research misconduct and can not impose the debarment. As such, the one-year debarment should be overturned.
Furthermore, I have categorically denied all the allegations. The allegations simply are not factual. We would be happy to provide you with our responses against the University and NSF’s Final determination.
Regarding the reproducibility of my work as a postdoc at CSU is concerned, as related to compounds 1-4 (from the report), the work has been reproduced by independent groups and also by researchers in Vivanco lab (where I am not a co-author) (at least 20 odd papers showed that catechin is toxic and is secreted by C. stoebe plants; see annex I).
Regarding your question related to the retraction of plant physiology paper, I strongly disagreed with the basis of retraction (which I specified in my email to Dr. Don Ort (Then, Chief Editor of Plant Physiology). The main area of disagreement with the retraction was the basis of the retraction cited by the corresponding author – which was completely irrational and non-scientific. Corresponding author claimed that Figure 1C-D and 2A-B is erroneous or non- reproducible. Figure 1C-D shows the chemical structure, whereas Figure 2A-B shows phytotoxic activity of catechin against Arabidopsis. There are at least 20 or so independent publications which established that both of these claims stand and independent labs have reported that catechin is not only secreted but is also phytotoxic against a range of plants.
The catechin work initiated by me during my postdoctoral term opened a new area wherein the phytotoxins involved in plant-plant interactions that could be further investigated for potential mode of action and genetic targets. My work showed that catechin phytotoxicity on the susceptible plants involve a specific signaling cascade to trigger the cell death and the North American plants are more susceptible to catechin toxicity compared to European counterparts. There were two different groups at CSU (Vivanco & Hufbauer) testing the hypothesis at both the lab and field scale that biological invasion in Centaurea is facilitated by secretion of phytotoxins. Dr. Hufbauer’ group has published that the concentrations reported for catechin secretion is towards the lower range then the original reports. Interestingly, Hufbauer’s group showed that catechin is secreted. The other claims of catechin secretion and phytotoxicity are all listed in annexure I. As per the retraction note, the data pertaining to Fig. 1C-D and 2A-B are erroneous or non-reproducible, contrastingly, both of these claims are shown independently in at least 20 other publications (see annexure I).
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, and sign up on our homepage for an email every time there’s a new post. Click here to review our Comments Policy.