University of Tennessee investigation finds manipulated images in Science paper

An investigation by the University of Tennessee Health Sciences Center, in Memphis, into a 2006 Science paper found evidence that three figures in the article had been manipulated.

Science sleuth Elisabeth Bik first flagged the paper, titled “Molecular Linkage Between the Kinase ATM and NF-κB Signaling in Response to Genotoxic Stimuli,” to the editors of Science in 2015. Today, Science issued an expression of concern for the paper: 

On 24 February 2006, Science published the Report “Molecular linkage between the kinase ATM and NF-κB signaling in response to genotoxic stimuli” by Z.-H. Wu et al.. Concerns about Figs. 4E and 4F and fig. S2A were posted to PubPeer in 2020. Because of the age of the paper, Science no longer had original files to examine. Science has now learned that relevant institutions have completed investigations and have concluded that there was figure manipulation. Science is contacting the institutions to obtain further clarification, but in the meantime we are notifying readers of our concern about the paper’s data integrity.

The paper has been cited 383 times, according to Clarivate Analytics’ Web of Science. At the time of its publication, all of the study authors were affiliated with the University of Wisconsin–Madison. The study’s first author, Zhaohui Wu, became an assistant professor at the University of Tennessee Health Science Center in August 2008.

In an email to Retraction Watch, Brian Fox, the University of Wisconsin–Madison’s research integrity officer, said that Shigeki Miyamoto, the study’s corresponding author and a professor at the university, was not aware of the fabricated images when the paper was submitted. Fox also mentioned that a separate institution had investigated the issue:

UW-Madison is aware that the University of Tennessee Health Sciences Center has determined that available evidence indicates it is more likely than not that fabrication of an image is contained in the Science paper you mentioned below. 

After examination of the available evidence, UW-Madison has concluded that Dr. Miyamoto had no knowledge of the fabrication. UW-Madison also supports Dr. Miyamoto’s efforts to work with the journal to correct the scientific record.

In an email, Miyamoto said that he “cannot make any comment at this time,” and Fox declined to address why the University of Tennessee was involved in the case. Mark Miller, the research integrity officer at the University of Tennessee Health Science Center, declined to comment on the investigation. 

Wu did not respond to our requests for comment.

The study’s manipulated images were originally spotted by Bik in April 2015. Bik says that she sent her concerns to the journal at that time. Then, nearly five years later, in March 2020, she posted about the manipulated images on PubPeer.

In an email to Retraction Watch, Bik explained the timeline of her discovery:

This paper is one I found during my 2014 / 2015 scan of 20K papers in 40 different journals. Of the 681 Science papers that I scanned, I found 9 papers with problematic images. I reported all of these to the journal six years ago, in April 2015. Only three of these papers have been corrected so far, and the other six were still unaddressed. A year ago, I posted these problems onto PubPeer, because I was disappointed that the journal appeared to have forgotten about my concerns. After I posted these on PubPeer, I had a good conversation with Dr. Thorp, the new Editor in Chief of Science since 2019, and asked him about the status of these investigations. I am happy that the journal, under his leadership, has started to address some of the remaining issues. 

Bik also explained the technical issues with the three different images:

In this particular paper, I noticed three potential problems, all duplications in protein blots. First, two adjacent lanes in the [NF-κB essential modulator, or “NEMO”] Western blot panel of Figure 4E appeared to be mirror images of each other. Second, two alpha-GST panels in Figure 4F appeared to be showing the same two lanes. The third problem was a potential duplication within a blot in Supplemental Figure S2A.

She also urged Science to retract the paper, writing:

Two of the three problems appeared to involve a duplication within the same photo, which is something that is very unlikely to have happened by accident. I am not sure why Science has decided to mark this paper with an Expression of Concern, and not with a retraction, especially since the affiliated university, the University of Wisconsin–Madison, has confirmed there was image manipulation. However, an EoC is a clear warning to readers that certain data in this paper is not trustworthy. But, it would have been even better if the journal had done that in 2015, after I reported my concerns. 

Bik continued:

An earlier EoC would have prevented many of these studies to have put their trust in this study.

I do hope that the Science editors and staff will continue their contact with the affiliated university, and change the EoC into a retraction. 

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].