Penn State prof earns second retraction, faces third following university probe

A professor of biomedical engineering at the Pennsylvania State University today lost a government-funded study in Science Advances, marking her second retraction. 

The researcher, Deborah Kelly, is also facing retraction of a paper in Current Opinion in Structural Biology after a review undertaken by her institution found “serious data integrity concerns” in the work, according to emails obtained by Retraction Watch. Kelly has hired a lawyer to fight the retraction, apparently without success. (Update on Sept. 12: The paper has now been retracted.)

Today’s retraction of “Structural analysis of BRCA1 reveals modification hotspot” cites “unresolved concerns in the integrity of the data presented,” including what appears to be alterations of cryogenic electron microscopy (cryo-EM) maps using an “eraser tool.” The study was funded in part through a National Institutes of Health (NIH) grant for US$353,386 to Kelly.

In a statement to Retraction Watch sent via her legal counsel, Kelly stood by her work. 

Due to “ongoing confidential proceedings” regarding the proposed retractions, “I consider substantive discussion about them to be premature at this juncture,” Kelly said. “At least one external review has very recently contradicted the conclusions of the journals as presented in the respective retraction notices.” 

“I remain confident that the ultimate results of these ongoing proceedings will confirm that the findings and data presented in these papers remain accurate, reliable, and scientifically sound.”

Concerns have been swirling around Kelly, a former president of the Microscopy Society of America, at least since early last year when a scientist at Newcastle University in the United Kingdom began flagging problems in her work on PubPeer. So far, Thomas McCorvie has pointed out issues with 21 of Kelly’s publications.

McCorvie said he first noticed problems with Kelly’s work when the now-retracted paper in Science Advances appeared in 2017. 

“At the time a colleague emailed Deb Kelly due to our concerns, but the conversation was not productive, and Kelly stopped replying,” McCorvie told us. “As far as I am aware Kelly did not acknowledge the erased regions in their two maps. This was all very suspicious, but we didn’t follow up on it at the time.”

Then a paper from 2022 in ChemBioChem rekindled McCorvie’s apprehensions. He also noticed further problems with the 2017 paper in Science Advances that made him suspect the cryo-EM maps had been “fabricated,” he said in an email. 

“Due to all these issues, I looked further into the Kelly lab’s work last year,” McCorvie explained. He said the 21 papers he has flagged on PubPeer:

are from the Kelly lab’s time at Penn State and Virginia Tech. There are also three papers with issues from Kelly’s time as a postdoc at Harvard medical school. The general concerns cover:

1. Reuses of figures and images (micrographs, orientation distribution plots) for different conditions and samples, 
2. Electron microscopy maps not having the expected pixel sizes,
3. Atomic models not correlating with the electron microscopy maps,
4. The electron microscopy maps not having the expected features at their stated resolutions.

There are many problems, and I believe it to be a deep-rooted issue. None of the authors have contacted me to explain why there are these problems.  

Kelly’s LinkedIn page says she is also the executive director of the new non-profit think tank Structural Oncology LLC.

The retraction notice in Science Advances, which follows an editorial expression of concern from March, stated:

After publication, readers raised concerns that the cryo-EM maps had been altered. As a particular example, the map for EMDB-8834 appeared to be altered using a volume eraser tool to create a spherical hole in the map. These concerns were presented to the authors and an editorial expression of concern was issued on the paper on March 13, 2024. The authors’ response was reviewed by multiple experts in Cryo-EM, and this review concluded that unresolved concerns in the integrity of the data presented in this paper remained. As a consequence, the editors of Science Advances have decided to retract this paper. Some but not all of the authors agree with this retraction, and others have not responded. Dr. Kelly does not agree with the retraction.

McCorvie’s PubPeer comments give a detailed account of the issues with the paper, including the apparent use of an eraser tool in two density maps. 

“Overall, I am incredibly concerned with the state of these maps. They have been clearly edited and their origin appears questionable,” McCorvie wrote in one comment, noting in another that he kept “finding inconsistencies” in the paper.

McCorvie added a caveat:

Any students involved in this paper should not be judged for the veracity of the findings, and maps. I am sure they are all very capable scientists. Responsibility of the edited maps however lies solely on the PI. This is clear as they are the common denominator across all flagged manuscripts and are the head of the lab. This also applies to other students in the Kelly Lab.

The new retraction comes just a month after the journal Nanoscale pulled a 2021 article titled “Microchip-based structure determination of low-molecular weight proteins using cryo-electron microscopy,” for “concerns about the reliability of the electron microscopy data.” (McCorvie described the problems here.)

Kelly was the senior and corresponding author, and opposed the retraction, according to the notice. 

“Independent experts were consulted who were not satisfied with the explanation provided by the authors, therefore this article is being retracted to protect the integrity and accuracy of the scientific record,” the notice stated.

Edward H. Egelman, a professor of biochemistry and molecular genetics at the University of Virginia in Charlottesville, was one of those independent experts. An editor for the Royal Society of Chemistry, which publishes Nanoscale, contacted him in July 2023 to ask for his assistance in evaluating the article, critiques of the data, and the authors’ response, according to emails seen by Retraction Watch. 

In Egelman’s assessment, “there are fundamental concerns about the paper.” 

Based on input from Egelman and another expert, the journal decided to retract the article, the editor informed Egelman in October. 

However, once the journal sent Kelly the proposed retraction, she responded with a rebuttal, which the journal asked Egelman to review in December. Kelly’s response did not change his mind. 

After some back and forth regarding raw data Egelman suggested the journal request from Kelly, the journal published an expression of concern in February “in order to alert readers that concerns have been raised regarding the accuracy” of some of the data in the article. “An investigation is underway,” the notice stated.

Frustrated that Nanoscale still had not retracted the paper, Egelman submitted a concern to the Committee on Publication Ethics (COPE) in March. In August, after the retraction was published, COPE informed Egelman it could no longer consider the concern due to “ongoing legal procedures.” 

The two retracted works have been cited 27 times in total, according to Clarivate’s Web of Science. 

Separately, in July Egelman published a detailed critique of another of Kelly’s papers that appeared in Advanced Materials. “Basic principles in mathematics and physics would need to be violated” to achieve the reported results, he wrote. 

McCorvie said Current Opinion in Structural Biology would soon be pulling another paper by Kelly and her colleagues – “Liquid-EM goes viral – visualizing structure and dynamics” – that he did not flag, and that all of the authors had been informed of the move. 

That article, which has been cited six times, was the subject of “a lengthy and comprehensive peer review process” undertaken by Penn State and “conducted by three independent experts from outside of Penn State,” according to a July 26, 2024, email from the school’s Office for Research Protections obtained by Retraction Watch.

The “review process concluded that there are serious data integrity concerns relating to volume and model” in the article, wrote Chief of Staff Courtney Karmelita. She added:

In addition, the corresponding author, when asked to produce original research data to justify the data, maps and models in question, did not do so. For these reasons, Penn State recommends that this paper be retracted. The corresponding author who is copied here (along with the other co-authors whose contact information I could find) declined to seek retraction despite having been urged to do so by the peer review panel, thus requiring that I, as Research Integrity Officer, reach out to you directly with this institutional recommendation for retraction.

A spokesperson for Elsevier, the publisher of Current Opinion in Structural Biology, confirmed they were investigating the article.

Egelman told us:

The investigation at Penn State concluded in May, and they refused to make any of their report public but passed the matter to ORI [the federal Office of Research Integrity] that they said would ultimately issue a report. Kelly was the Chair of an NIH study section as well as the recipient of funds for many years from [the National Cancer Institute], so there is potential embarrassment for many parties.

In a statement, Penn State said it had “conducted a peer review with external subject matter experts” which “confirmed the presence of unreliable data in several papers.” The university contacted journals to request retraction of the papers and reported its findings to ORI, the statement confirmed.  

A spokesperson for ORI said the office was “not able to confirm or deny the existence of any potential pending cases.” 

In a letter to Current Opinion in Structural Biology on Kelly’s behalf, Paul Thaler, a partner at the Washington, D.C. firm Cohen Seglias Pallas Greenhall & Furman, disputed the retraction. Thaler argued the editors and university investigation had not provided Kelly with sufficient evidence the data in the paper were unreliable.

Thaler copied another lawyer, Pittsburgh, Penn.-based Heather S. Heidelbaugh of Leech Tishman. (In 2020, Heidelbaugh ran for attorney general of Pennsylvania against incumbent Josh Shapiro (now governor of the state) and lost.)

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].