The Chinese education ministry is investigating scientific articles authored by high-profile immunologist and university president Cao Xuetao, following suggestions that dozens of papers contain potentially problematic images. Four journals also say they are examining papers from Cao.
The scrutiny comes after US microbiologist Elisabeth Bik raised concerns two weeks ago on Twitter and the post-publication peer-discussion site PubPeer about images in papers written by Cao and his group.
Cao is the president of Nankai University in Tianjin, and his team has pioneered the development of cancer immunotherapies in China. He says that his group is investigating the papers in question, and he is confident that the issues raised do not affect the papers’ conclusions. Cao has been a prominent voice for strengthening research integrity in China, and gave a speech on the topic at the prestigious Great Hall of the People in Beijing earlier this month.
Bik has flagged potential problems in about 50 papers co-authored by Cao on PubPeer, and other users, most anonymous, have raised similar issues concerning another handful of papers from the group. As of 27 November, images in 63 papers that the team published in 28 journals since 2003 have been flagged on the site.
In some papers, Bik says, seemingly identical images are labelled as representing different biomedical experiments. In others, features such as patterns of dots that represent biological data appear “unexpectedly” duplicated in the same image, she says.
“That would be the equivalent of someone showing you a photo of the night sky, and you would see two Big Dipper constellations in the same photo,” says Bik, who has developed a reputation for spotting and raising potential problems in scientific images and figures.
During a press conference on 22 November, Xu Mei, a spokesperson from China’s Ministry of Education, said the ministry is investigating the articles in question and the “relevant” institutions. Cao is also the director of the Institute of Immunology at the Second Military Medical University in Shanghai, also known as the National Key Laboratory of Medical Immunology, and most of the articles in question list this affiliation.
Bik told Nature that she cannot comment on whether the issues are the result of research misconduct. “It is up to the affiliated institutions to investigate and conclude,” she says. Although Cao’s name is on the papers, often as the corresponding author, it is not clear how closely he was involved in the work.
On 17 November, Cao responded on PubPeer to Bik’s comments, saying that his team and collaborators have made it their priority to re-examine the identified manuscripts, raw data and lab records. “We’ll work with the relevant journal editorial office(s) immediately if our investigation indicates any risk to the highest degree of accuracy of the published records,” he wrote.
He also said he is confident that the conclusions in those papers remain valid and the work reproducible. He apologized for “any oversight on my part” in his role as a mentor, supervisor and lab leader, and added that there is no excuse for a lapse in his supervision or leadership. “I’ll use this as an invaluable learning opportunity to do better not only in advancing science, but also in safeguarding the accuracy and integrity of science,” he wrote.
Cao did not respond to requests for comment on the issues raised about his team’s papers on PubPeer. Nankai University directed Nature to Cao’s statement on PubPeer.
Individuals, including some who seem to be Cao’s co-authors, have responded on PubPeer to some of Bik’s queries. In at least one case, a co-author acknowledges that the wrong photograph has been published. In another case, commentors suggest that images flagged as duplicates by Bik were in fact pictures of the same cells taken over time, but that the figure’s labels were unclear. The explanations given in those instances have been satisfactory, says Bik.
In comments about a few other papers, Bik questions images that the authors have already acknowledged in published errata.
But the authors have not yet responded to questions raised about other papers, in which features such as bars or patterns of dots appear multiple times in the same figure or image, she says.
Several researchers who have not collaborated with Cao or Bik have told Nature that the figures she has flagged seem suspicious. Nicole La Gruta, a molecular biologist at Monash University in Australia, says that in her opinion: “It is clear from the multiple images that I have seen that these are definitely manipulated.”
Wouter Masselink, a postdoctoral molecular biologist at the Vienna BioCenter in Austria, agrees that some of the images require explanation. “I hope the institutions and universities that Cao is associated with launch a formal and independent investigation to find out how and where these artefacts ended up in the published manuscripts,” he says.
Bik says she plans to contact the journals that published the papers that she has identified. But the comments on Twitter and PubPeer have already caught the attention of some of the 28 journals concerned. Representative from four — Science, Nature Communications, Cardiovascular Research and Molecular Immunology — told Nature that they had heard about the potentially problematic papers in their journals and were reviewing them.
Elisa De Ranieri, the editor-in-chief of Nature Communications in London, says the journal saw posts on Twitter and PubPeer that raised issues over potential image manipulation. "We are now going to look into this matter for what concerns any relevant Nature Communications papers, as per our usual research-integrity processes," says Ranieri.
Meagan Phelan, a spokesperson for Science’s publisher, the American Association for the Advancement of Science in Washington DC, says Science is also reviewing an article in the journal that Bik flagged. She added that it's up to institutions to investigate any possible misconduct, which would inform any decisions the journal made.
A spokesperson for Germany’s Academy of Sciences Leopoldina, of which Cao is a member, told Nature that the academy “will observe the further investigation of the matter closely”.
Cao received a Nature Award in 2015 for excellence in mentoring, and he is co-editor-in-chief of Cellular & Molecular Immunology, a journal published by Springer Nature. A spokesperson for the company says it does not appoint the journal’s editorial committee. “We have just become aware of the concerns that have been raised around images in some papers co-authored by Professor Cao. We have no further comment to make at this stage,” they say.
On 22 November, Nature Immunology posted an ‘Editor's Note’ on two of Cao’s papers identified by Bik. One note says the authors had flagged a duplicated image before publication but it was not corrected in time; in the other, the journal says a duplicated image was “inadvertently introduced during the production process”.
A spokesperson for the journal says it is aware of the concerns about Cao's papers and has no further comment to make at the moment. “It is important to note that issues with images are often not a result of research misconduct, but they should be assessed carefully and appropriate action to correct the scientific record taken as necessary,” they said. (Nature’s news and comment team is editorially independent of its publisher, Springer Nature, and of other Nature-branded journals.)
Image duplication has become a persistent problem in the scientific literature. In 2016, Bik and colleagues scanned more than 20,000 biomedical articles published in 40 journals between 1995 and 2014, and found that one in 25 articles contained inappropriately duplicated images. Bik says that she reported 782 suspect papers to journals in 2014 and 2015, which has resulted in almost 200 corrections and more than 50 retractions or expressions of concern, she says.
Bik says that about two weeks ago, she started following up the outcomes of the papers that she had flagged to journals in 2014 and 2015, and noticed that three papers with multiple figure issues had come from the same lab — Cao’s. “I thought this was a group that’s worth further investigating,” she says.
She started posting suspicious images on PubPeer and Twitter, not knowing, she says, that Cao is a highly respected research leader in China. She says that she has now scanned 325 papers co-authored by him.
She didn’t expect that her posts about Cao would prompt so much attention, she says. As well as encouraging messages about her latest findings, Bik has received tips about potential image issues in other papers from Cao’s team. She has also received some disparaging tweets, but she says that hasn’t deterred her. “I’m not here to make friends,” she says.
Nature 576, 16-17 (2019)