Exclusive: Ohio State researcher kept six-figure job for more than a year after a misconduct finding

In 2016, Mingjun Zhang, a biomedical engineering researcher at The Ohio State University, along with collaborators, published a paper that explored the mechanism behind ivy’s impressive adhesive strength. In it, the authors claimed to report the genetic sequences of the proteins making up the adhesive.

The paper, entitled “Nanospherical arabinogalactan proteins are a key component of the high-strength adhesive secreted by English ivy” appeared in PNAS on June 7 and attracted some media attention.

But shortly after publication, an anonymous whistleblower sent a letter to OSU and PNAS simultaneously: “The authors have knowlingly [sic], intentionally, repeatedly, and substantially misrepresented data in order to publish the manuscript.”

An internal investigation lasting until July 2018 found that Zhang had committed misconduct, according to nearly 200 pages of documents and emails obtained by Retraction Watch through a public records request. But Zhang, who joined OSU in 2014 and denies all allegations of misconduct, remained in his job — which paid more than $170,000 per year, according to records kept by the Buckeye Institute — through 2019. And the PNAS paper has yet to be retracted, two years after OSU requested that the journal do so.

The study in question — which has been cited by 19 other articles, according to Clarivate Analytics’ Web of Science — looked at the nanoparticles in ivy’s sticky residue. The authors argued that these residues were primarily composed of arabinogalactan proteins (AGPs) which are common in plant secretions. They claim they were able to determine both that these nanoparticles were “spheroidal” in shape and show they consist of AGPs by genetically sequencing the particles. They also tried to synthesize a “biomimetic” ivy residue from similar components and compare its mechanical properties to those of regular ivy residue.

The paper — funded primarily by the Army Research Office and the National Science Foundation — purportedly found “that the ivy nanoparticles are predominantly composed of AGPs.” Zhang said in a May, 2016 press release about the study that ivy adhesives could be useful in developing armor coatings, among other things: “By understanding the proteins that give rise to ivy’s strength, we can give rise to approaches to engineer new bio-inspired adhesives for medical and industry products.”

It was a quote that would later be used against him by members of one of the OSU committees that investigated his work.

On July 1, 2016 – three weeks after Zhang’s paper was published – the anonymous whistleblower’s letter to OSU was forwarded to Jennifer Yucel, then the director of the research compliance office at OSU, according to an initial report. On July 12, she met with Richard Hart, chair of the department of Biomedical Engineering, and Randy Moses, associate dean for research at the College of Engineering to go over the allegations against Zhang and “conduct a Preliminary Assessment.”

They decided they needed someone with relevant expertise, so the next day they enlisted the aid of Michael Ibba, then the chair of the microbiology department at OSU, because of his “extensive” experience identifying and characterizing new genes. 

The initial report said, on July 15, Etta Kavanagh, then editorial manager at PNAS, reached out to OSU regarding a whistleblower’s complaint. 

The next day, Ibba wrote to Yucel saying: “In my opinion, the case as presented does not constitute research misconduct, but does nevertheless raise concerns related to data interpretation by Dr. Zhang.” At this point, he wrote, it wasn’t clear whether Zhang was out of his depth or deliberately misleading the journal. Five days later, Yucel replied to Kavanagh that she had determined that the letter was the same as that sent to OSU. 

On July 25 and 26, 2016, Kavanagh sent Yucel reviewer letters and the author’s responses. On July 26, the investigators met to go over the documents, Ibba’s assessment, and “some basic sequence and homology analyses conducted by Dr. Yucel,” the report said. They decided that there was enough credible evidence of possible misconduct to go forward with an inquiry.

August 23, Yucel, Moses, and Hart met with Zhang for the first time to tell him of the allegations against him and to “determine who had participated in the generation of the data in question,” the initial report said. “Immediately” after the meeting Yucel went to Zhang’s lab to gather the relevant research records and “and Dr. Yucel briefly met with [redacted],” the report said.

After getting the approval of Caroline Whitacre, the senior vice president for research at OSU, they moved on to the next step – forming a Committee of Initial Inquiry which met for the first time on Oct. 11, 2016. 

The committee first formally interviewed Zhang on Jan. 11, 2017. The members inquired into Zhang’s general lab routines and asked why he believed he had correctly identified the sample in the paper. Over the following three months Zhang and his attorney, Emily Haynes, provided letters of support and other materials which the committee reviewed.

On April 28, 2017, the committee sent their determination to Whitacre. They had voted three to one that “that the allegations did not have sufficient substance to indicate possible Research Misconduct and therefore should be dismissed,” according to the initial report (bolding from the report).

But the case was not closed.

When the committee sent Whitacre their conclusions, she asked them to reevaluate based on comments they had received from Zhang’s attorney on April 5. This time, the report said, the committee members voted 3 to 1 “that the allegations did have sufficient substance to indicate possible Research Misconduct and warrant investigation…” (bolding from the report)

What had changed was that Zhang’s interviews and records started to diverge. In the preliminary interviews, Zhang suggested he kept the work at arm’s length. Though the name is redacted, he blamed someone else, likely the paper’s first author who was then a graduate student in his lab, Yujian Huang. “Dr. Zhang consistently asserted that the decisions in the research were made [redacted] and that Dr. Zhang did not have close involvement.” 

Zhang told the committee that “as the paper was continually resubmitted to journals, he became less passionate about this manuscript as other matters were becoming more pressing,” the report said. “Dr. Zhang states that at this point, he did not even read the paper.” He said he “did not do the experiments or prepare the manuscript, or prepare or submit the response to reviews.”    

The review committee was concerned by Zhang’s apparent lack of involvement with the research going on in his lab, but initially took this to mean Zhang was, at worst, being grossly negligent in his responsibilities as a principal investigator, not as evidence of misconduct or data fabrication.

But letters from his former and current mentees and trainees gathered by Zhang told a different story. Those letters argued “that he was closely involved in oversight of their projects,” the report said, and “that he is closely involved in mentoring.” The apparent contradiction led the committee — which also noted that many of the outside letters were from people with little relevant expertise in the subject matter, and that two of the letters were identical — to become suspicious of Zhang’s claims that he was barely involved in the work. 

“Dr. Zhang has tried to present himself to Dr. Yucel and this committee as a detached PI, who does not have the expertise with respect to the specific allegation,” the committee wrote, but the letters suggest that “Dr. Zhang is engaged in all aspects of the process, and fully understands the implication/questions of DNA sequence homology.” Documents from Zhang’s computer – details of his back-and-forth discussions with peer reviewers – also showed how closely he was involved with the article in question, according to the report.

The conflicting statements led the committee to doubt Zhang’s credibility and suspect deliberate fraud. That prompted them to change their vote, which in turn led OSU to form a College of Engineering committee that continued the investigation.

In Zhang’s response to the first committee’s preliminary findings, he said he chose the ivy sequence he presented and discarded the others from McLab because they didn’t have “adequate proline residues,” according to the report, which said the residues are a defining feature of the arabinogalactan proteins which Zhang  anticipated finding – implying he cherry picked the results. “This raises questions about whether only data consistent with the AGP hypothesis were accepted or were considered, and whether data were selected based on their consistency with the hypothesis,” the committee noted. 

Li Tan of the University of Georgia, a co-author on the paper, told Retraction Watch that he joined Zhang’s collaboration late, after the genetic data was taken and Zhang had submitted the manuscript to several journals. He said he did some chemical analyses and contributed to several figures in the supplemental information. “As an outsider, I do not know how Dr. Zhang handled the genetic part of his research,” he wrote in an email.

The sequence data should have raised red flags among PNAS editors, according to the committee. “The Editor should have required the authors to provide clear and convincing experimental evidence of the identity of the clone…” Reviewers did have concerns. Reviewer 1 at PNAS, for example, wrote that the genetic sequence of the ivy particles “does not resemble any known” arabinogalactan protein, and that they look more like a “cytochrome oxidase subunit.”

The whistleblower claimed to have reviewed the work for Nature Communications back in July, 2015, and had similar concerns. At the time, the whistleblower wrote, “I could not rule out incompetence as an alternative explanation.” But the whistleblower wrote of one of the author’s findings — “that Big-PI predicts a GPI anchor site for this protein” — that it was “not possible to make an honest mistake about such a claim.”

Nature Communications’ reviewer 4 wrote, according to the subsequent report by an OSU College of Engineering committee: “Unfortunately, as I studied the experiments in this paper to understand fully the evidence, it became clear to me that many of them are very poorly conducted, and very much over interpreted. The experiments are sometimes not as they are described, they are redundant, or they are uninformative…”

The reviewer said of the figure showing the alleged genetic sequence — the central finding: “There being almost nothing correct in this figure, I searched the amino acid sequence against NCBI. It is identical to a fungal cytochrome C oxidase subunit. The authors have not identified an ivy AGP. They have cloned a fungal protein.”

The July 31, 2018, College of Engineering investigation report also pointed out that while Zhang indicated in interviews “that he believed that the sequence of the protein was not important or critical to the manuscript,” his quotes in the May 2016 press release suggested that he understood the sequence’s importance. 

The report concluded by a 2 to 1 vote that Zhang “recklessly falsified and misrepresented” the genetic data used in the PNAS paper, which “does constitute Research Misconduct.” The committee recommended three sanctions against him: Zhang’s new students going forward the next three years would have to work with a co-adviser, Zhang himself would have to complete the “CITI Responsible Conduct of Research training” within two weeks of the final report’s completion, and Zhang would need to include “experienced, qualified domain-specific co-Principal Investigators on all multi-disciplinary grant proposals” outside his area of expertise for three years. 

But Zhang remained in his job until he resigned at the end of 2019, according to a university spokesperson.

In September 2018, two months after the final report of the College of Engineering committee was completed, OSU held a summit in which it pledged to make research integrity a priority. The summit came seven months after Susan Garfinkel, formerly of the U.S. Office of Research Integrity, joined the university as assistant vice president for research compliance.

That appointment came toward the end of a decade during which OSU saw itself in the news numerous times for allegations of scientific misconduct by its faculty. In 2012, Terry Elton was found to have manipulated dozens of figures, in what OSU initially said was just a case of disorganization. In 2016, the university stripped a former student of a PhD, two years before he would be found by the ORI to have committed misconduct.

In 2017, The New York Times reported that allegations of misconduct by Carlo Croce — whose name should be quite familiar to Retraction Watch readers — had been investigated multiple times. That same year, exercise researcher Steven Devor resigned his post amidst a lawsuit brought by Crossfit alleging that he had fudged his data. (Crossfit has prevailed.) And also the same year, OSU revoked Jodi Whitaker’s PhD for misconduct — after asking whistleblowers’ institutions to sanction them for making the university’s deliberations public.

In 2018, the university went public with findings of misconduct by cancer researcher Ching-Shih Chen, and a former cancer researcher, Samson Jacob, earned a slew of retractions.

The College of Engineering committee also recommended “immediate” retraction of Zhang’s 2016 paper along with the removal of his three GenBank entries for ivy’s genetic code. At time of writing all three genes have been removed from GenBank, but the article remains online.

On September 10, 2019, PNAS added an expression of concern to the piece. But it bears no mention of misconduct or the many other falsehoods in the text: “The editors wish to note that they have been informed of a concern regarding the data and text related to Fig. 3 and Fig. S4A. The editors are issuing an expression of concern to alert readers while we await information that will allow us to determine whether further action is required.”

Meanwhile, most of the 19 articles that have cited Zhang’s study cited it after OSU sent PNAS their retraction request. Ten cited it after the expression of concern was posted, and two of the 19 articles are considered “Highly Cited Papers” by Clarivate Analytics’ Web of Science. 

Two years after OSU’s Yucel sent PNAS a request for retraction — included in the investigation documents — and more than four and a half years since they received the whistleblower’s allegations, the paper remains unretracted. May Berenbaum, editor in chief of PNAS, told Retraction Watch: “For PNAS, the matter remains unresolved because it’s still under review by the National Science Foundation and we’re awaiting the outcome of their inquiry before taking further action on the paper. We haven’t looked into additional manuscripts from Mingjun Zhang or coauthors because the initial case is still pending with NSF.”

In its July 2018 report, the College of Engineering committee notes that Zhang apparently did accept that the data were falsified, but did not accept that that falsification was misconduct: “Dr. Zhang did not dispute the finding that the data in question under allegation #1 were falsified; that while Dr. Zhang did not agree with the finding of Research Misconduct, Dr. Zhang did agree to accept the proposed sanctions…”

Indeed, Zhang told Retraction Watch, “I don’t have a research misconduct story.”

“I entirely disagree with any and all allegations that I engaged in any research misconduct,” he wrote in an email copying his attorney, Paul Thaler of Cohen Seglias. “To date, there have been no such findings by NSF. I have always endeavored to be a responsible researcher who maintains strong records. I conduct all due diligence to ensure my interdisciplinary research with any and all coauthors is conducted in a collaborative, accurate manner consistent with the scientific community’s expectations. The areas of concern in the allegations have no impact on the paper’s major contributions and conclusions.”

Leming Sun, a co-author on the paper at Northwestern Polytechnic University, told Retraction Watch that he did not believe Zhang committed misconduct, and that Huang was responsible for the data.

Retraction Watch called a phone number previously used by Huang to find that it is now disconnected. Huang did not reply to multiple email requests for comment.

“Ohio State rules state ‘Sanctions shall be commensurate with the severity of the research misconduct,’” the university spokesperson said. “As the records show, the investigating committee weighed numerous factors specific to the case to determine the appropriate actions for the misconduct found.”

Several days after Retraction Watch reached out to OSU for comment, Zhang’s lab website, which had been live until then, began producing a 404 error.

Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].