France’s national research council has ruled that one of its plant biologists committed misconduct through manipulation and data fabrication in published figures, but it cleared another researcher whom it had heavily sanctioned in 2015.

The ruling should add some clarity and closure to the long-running saga, though the cleared researcher, Olivier Voinnet, is now raising fresh questions over how the French research agency, CNRS, handled its initial investigation.

The CNRS announced its conclusions this week, following a fresh inquiry led by the agency — with the participation of the Swiss Federal Institute of Technology Zurich (ETH Zurich) — into five articles published by researchers at a now-defunct lab at the CNRS Institute of Plant Molecular Biology in Strasbourg, France. The lab was renowned for its work on a gene-silencing technique called RNA interference.

The CNRS and ETH Zurich each drew their own conclusions about their respective staff members, on the basis of the inquiry’s report.

ETH Zurich released its conclusions last month, saying that the inquiry found “severe” and “intentional” manipulation of research figures. But it said that Voinnet, a former leader of the Strasbourg lab and a prominent CNRS scientist who has been on secondment to ETH Zurich since 2010, “did not perform, order or scientifically endorse such manipulation”.

However, ETH Zurich concluded that, as group leader and a co-author of four of the papers, Voinnet also bore overall management responsibility; the institution therefore extended until 2023 an existing probation it had implemented after its original 2015 investigation. Conditions of his probation include having his publication activity monitored and having a mentor assigned.

Second opinion

The CNRS has now reached a similar conclusion to ETH Zurich with respect to Voinnet. Nature has obtained a copy of the conclusions of the meeting of the CNRS disciplinary committee which advises CNRS management on appropriate sanctions. The meeting took place on 10 July 2018 to consider the situation concerning Voinnet. It states that after studying the CNRS-ETH Zurich report, and after interviewing the inquiry's president and Voinnet, it found no evidence of any serious wrongdoing by Voinnet – and subsequently voted 7 to 0 (with one abstention) in favour of no sanctions against him.

In a statement released on 3 October, CNRS reiterated that its disciplinary committee had found no evidence that Voinnet was responsible for the fabrication of figures in the investigated papers. But it said that as a former head of the group, Voinnet bore some responsibility and so gave a reprimand which will expire in three years’ time, and which, in legal terms, is category 1, the least serious sanction under the four-tier scale of severity used by the French civil service.

The finding contrasts with the results of the CNRS investigation in 2015, which found Voinnet guilty of research misconduct and suspended him from the organization for 2 years — a category 3 sanction. At that time, the agency found no evidence of data fabrication, but said that the intentional manipulation of figures in the papers breached accepted ethical standards for putting together figures to show selections of raw data.

CNRS has also now reported that the joint CNRS-ETH Zurich report concluded that another former researcher at the laboratory, Patrice Dunoyer, committed misconduct in the figure manipulations — and in corrections to the manipulated papers — including fabrication of data. The CNRS said Dunoyer would receive the sanction of a demotion, which is a more severe punishment than Voinnet’s, but still a relatively low, category 2 sanction.

In 2015 Dunoyer received a 12-month exclusion from CNRS without pay for scientific misconduct, with 11 months of these served as a suspended sentence.

Alain Schuhl, deputy director-general in charge of scientific affairs at the CNRS, told Nature that the latest ruling also means that this suspended sentence will now kick into effect.

The official bulletin of the CNRS published on 9 October confirms the charges and sanctions against Dunoyer. Consistent with the conclusions of the 10 July 2018 meeting of the CNRS disciplinary committee, it makes no mention of any wrongdoing by Voinnet, or of the minor sanction given to him by CNRS.

Dunoyer has so far not replied to Nature’s requests for comment. Loic Dusseau, the lawyer representing Dunoyer, told Nature that he had been unable to comment until now because, beyond the 3 October press release, Dunoyer had not formally been informed of the details and grounds of the sanctions against him. Dusseau says that he only learnt of the details of official bulletin this morning, and has arranged to speak to Dunoyer.

The latest probe was instigated at the initiative of Voinnet, according to ETH Zurich, after he carried out his own in-depth investigation in summer 2016 into the group’s papers and raised the possibility of more-serious data fabrication than had been found in 2015.

Voinnet says that “CNRS’s reprimand is perfectly in line with the conclusions of ETH Zurich last month exonerating me”. But he takes issue with CNRS's 2015 ruling: “A reprimand is what I only should have got in 2015 — and not a two-year suspension.”

Schuhl declined to comment on the seeming reversal of responsibilities in the CNRS's latest conclusions and those of its 2015 investigation. He argued that the matter of its 2015 investigation is closed.

Case timeline

Questions about papers co-authored by Voinnet were first raised in January 2015 on the PubPeer website — a forum that allows anonymous commenting about research articles, and which has often exposed misconduct.

The CNRS announced in April 2015 that it had set up a commission of high-level experts to investigate the affair— and in July that year it announced the sanctions against Voinnet.

The announcement of sanctions mentioned in the lengthy official bulletin of the CNRS at the time refers only to an inquiry commissioned in early 2015 , and to an 8 June 2015 meeting of CNRS's disciplinary committee.

Nature has obtained a copy of the confidential report of this inquiry, which is dated 4 February 2015. It shows that the inquiry comprised three CNRS scientists and two from other French research organizations that it met between 29 January and 2 February 2015, and that, during this time, the experts inspected the relevant articles and interviewed several of those concerned, including Voinnet and Dunoyer.

The February 2015 report makes no mention of lab notebooks, raw data and the like — as the latest investigation has done —and runs to just four pages.

Voinnet hopes that the recent investigation will have the beneficial effect “of lifting the cloud of suspicion that has hung over many other members of the lab”, he says.

But he told Nature that he now intends to take administrative legal action against the CNRS to challenge the grounds on which the agency based his 2015 sanction.

In response, Julien Guillaume a spokesman for CNRS said that the latest joint CNRS-ETH Zurich investigation has no bearing on the sanctions pronounced in 2015 – an argument that Voinnet rejects.