The saga that led to the resignation of neuroscientist Marc Tessier-Lavigne from the presidency of Stanford University offers many lessons about scientific research integrity, the behaviour of research institutions and academic journals, and conflicts of interest, the power of journalism and human psychology.
There was public knowledge within the post-publication review community that there were problems with images in some of Tessier-Lavigne’s articles as far back as 2015, around the time that he was being considered for the Stanford University leadership. To his credit, he contacted the journals where some of the suspect data was published and attempted to address some of the issues through corrections of the articles. However, he relied upon a flawed exculpatory excuse for the origin of the image duplication and failed to insist on the publication of the corrections when the journals did not follow through.
The images in question were mainly generated through experiments where proteins were separated in an electric field and identified and quantified by binding to an antibody that was specific to them. The product of the procedure is called an immunoblot or Western blot, and the image displays black “bands” (on a background) whose intensity is reflective of the amount of the protein(s) of interest in the sample.
One journal declined to take any action, because the image problem that had been originally identified was based on a practice that was accepted at the time of publication but is no longer allowed. Troublingly, the journal was also aware of the duplication of panels that purported to represent the results of two different experiments. There were no visible bands, meaning that there was no detection of the proteins of interest, but the images in the “empty” panels should not have been identical; they do, in fact, display the same background and could therefore not have originated from distinct experiments. The journal gravely erred by not deeming the duplication as a serious enough matter about which to notify readers, although this approach is frustratingly common. The other journal simply forgot about making the corrections.
First lessons: it’s best to confront issues as they arise, but one has to persist in fixing them. Second, scientific journals will frequently do the bare minimum or nothing when presented with evidence of violations of academic standards. Third, researchers are so enamoured of their publications and unwilling to admit to serious flaws that they will grasp at erroneous rationalisations that justify their colleagues and themselves. Finally, images purporting to show “nothing” are just as much important data as those that show “something” and duplication of “empty” panels or “empty” areas of panels (“background”) should be treated as illegitimate practices. They are likely to be indicative of other malfeasance. Data from “control” experiments are real data and should be regarded as such.
From the public perspective, the troubles with the articles with Tessier-Lavigne as an author percolated to the surface in late 2022 as the culmination of the reinvestigation by bloggers and scientific sleuths of the papers in question as well as others. A student journalist at the Stanford Daily became aware of the controversy and wrote a series of articles about the matter. The initial response of the Stanford board of trustees, to whom the president reports, was to appoint a subcommittee consisting of some of its members, to review the issues relating to the Tessier-Lavigne articles. Egregiously, one of the members of this small “special” committee was a founder of an investment firm that possessed an $18 million (£14 million) stake in a company founded by Tessier-Lavigne. It took an inquiry from the Stanford Daily for that trustee to recognise that “question of potential conflicts”.
Next lessons: If one doesn’t firmly address serious allegations about research conduct, they might appear to subside for a while, but they are likely to resurface. Second, it often requires the intervention of the media to shine a light on flawed scientific articles for the issues with them to be addressed. In my own experience, it was not until after the New York Times made inquiries and published an article that included some of my findings of violations of scientific norms in an ensemble of papers that most of the journals began to issue corrections and retractions. Finally, there is a staggering blindness of American university boards of trustees – aka wealthy donors – to conflicts of interest.
Stanford’s trustee board belatedly realised that the only credible investigation that could be conducted would have to be executed by outside experts. A highly distinguished scientific panel was established. They were aided by technical consultants with expertise in research integrity and forensics. Although the investigation by the panel was limited to those Tessier-Lavigne articles about which allegations had been made, it thoroughly considered the images in question and, through the forensic analyses, identified some additional issues. The work of the panel validated virtually all of the post-publication review concerns.
Next lessons: Once consideration of a research integrity violation has passed the stage of a preliminary inquiry to evaluate whether there is any possible validity to the allegations, outside experts must be appointed as investigators. Virtually every internal inspector will have a conflict of interest. Second, both scientific and forensic expertise is necessary – the latter was particularly displayed to advantage in the published results of the investigation. Finally, the community of scientific detectives, identified, pseudonymous and anonymous, should be acknowledged for their efforts.
Further lessons: One can achieve academic success for some time despite having engaged in behaviour that violates scientific norms - however, perhaps not enduringly.
Three of the problematic articles found to display significant image manipulation and that are recommended for retraction share a common author beside Tessier-Lavigne (unnamed in the report). This author, subsequent to her experience as a postdoctoral researcher in the Tessier-Lavigne laboratory, became an assistant professor at a leading US university, undoubtedly, at least in part, because of the publication of the suspect articles. She was subsequently promoted to associate professor.
Although it is unclear in these cases who was at fault, problems that emerge with data in an accepted manuscript need to be communicated expeditiously with all concerned, including the publishing journal.
Another big question for consideration has also emerged: What is the responsibility of researchers for data that are published in an article who are not authors and what are the consequences for them when they are found to be flawed? I do not believe there is a consensus on this matter.
There are four levels of culpability when one is considering research misconduct.
- Knowingly fabricating or falsifying data or engaging in plagiarism. Being aware that such an act has occurred prior to publication is equally blameable in an author.
- Becoming aware after publication that there has been evident fabrication, falsification or plagiarism and denying that those violations of norms have occurred.
- Becoming aware after publication that there has been fabrication, falsification or plagiarism and, while admitting the occurrence, grasping for some implausible explanation or minimising the significance by claiming, for example, that the experiment was “just a control experiment” or that the falsified results have been “verified” by someone else.
- Incompletely addressing identified serious problems.
Of course, there are nuances when one is distinguishing between intentional misbehaviour and honest errors. Nevertheless, in the case of Tessier-Lavigne, it is fair to conclude that his culpability, according to conclusions of the Stanford-assembled scientific panel, is low on the scale – somewhere between levels 3 and 4.
Ultimately, Tessier-Lavigne was not found guilty of misconduct, but his management and oversight of his laboratory was found wanting. Facing no serious penalties, he will be able to further his substantial scientific achievements at Stanford.
However, it was precisely his qualities as a leader that came into question, and the deficiencies that were identified undermined his ability to continue as president. The prominence of his role at a distinguished university brought considerable attention to this incident. Despite these particulars, everyone engaged in the research endeavour can learn from the example of his experience.
David Sanders is associate professor in biological sciences at Purdue University.
