Every good cheater knows this: to avoid getting caught, you have to reinterpret, paying particular attention to what you copy. Because two productions with the same absurdities, that catches the eye. This is precisely what betrayed a vast network of falsifier biologists, which researchers have just brought to light, in a scientific article published on May 15 in the journal International Journal of Cancer.
The authors report having discovered a series of inconsistencies, repeated from article to article, suggesting the mistakes made by cheaters who are unaware of what they are copying. “Everything indicated that the scientists had made errors in their copying, when they had not simply invented research,” explains Jennifer Byrne, the biologist at the origin of these revelations, professor of molecular oncology at the University from Sydney.
In total, 420 papers were identified by this specialist, one of the 10 most influential scientists in the world according to the prestigious journal Nature. Most were written by Chinese scientists and published in unscrupulous journals, but some still managed to pass the review of major journals, such as Cancer Letters Or Oncogene. The incriminated researchers reported having dissected tumor cells, which according to specialists… do not exist.
In 235 cases, errors were repeated throughout the work. After verification, no mention of these cells with strange names was found in any register, except in these dubious articles. The sign, according to Jennifer Byrne, that certain reported experiments were simply invented: “I don’t see why serious researchers would have gone to great lengths to prepare, dissect, modify existing cells, then suddenly, would have started to invent some of them. But I don’t have proof of it. Still, as it stands, it is difficult to trust this work,” explains the specialist.
Imaginary genetic sequences
The errors were found more specifically in the mentions relating to the genetic sequences of very specific cells, commonly used in oncology. This involves samples of tumors, cultured in the laboratory. Scientists use these cells for all kinds of tests because they are simple to manipulate. Their barbaric name, a series of letters corresponding to their DNA, is only understandable to the most discerning connoisseurs like Jennifer Byrne. Otherwise, it is impossible to decipher them, which explains why the frauds have been able to persist in the literature without attracting attention.
Jennifer Byrne is not her first investigation. In 2017, she had already uncovered around forty articles with strange similarities, especially in their inconsistencies. Again, all from China. At the end of this previous investigation, published in the form of a scientific article in the journal Scientometrics, 17 articles were retracted, that is, unpublished and considered unusable.
The affair had “horrified” her. Since then, she has specialized in this type of investigation. Especially since these cases are relatively numerous, due to the proliferation of poor quality newspapers in recent years. His analyses, boosted by a homemade algorithm developed with French researcher Cyril Labbé, went so far as to find 19% false in certain samples: “I had the impression of swimming in an ocean of false science, I no longer saw “that”, she panics. In certain articles, researchers attribute functions to cells that they do not have. In others, their genetic sequence is wrong.
This work does not allow us to deduce the proportion of forgeries – they are not representative – but they show that there is an urgent need to address the problem: “These are needles in haystacks, but if we let it happen, soon the entire field will be invaded”, explains Guillaume Cabanac, one of the authors of the study, a computer science researcher specializing in the analysis of scientific corpora, also considered one of the most influential scientists on the planet, according to Nature.
Avoid the colossus with feet of clay
Most of the time, these false cells are not enough to prevent advances in cancerology, because they are mentioned in very specific experiments. But they can waste time. Or pollute serious work, simply because in science, we cite a lot of what others produce: “In the end, we find ourselves with a colossus with feet of clay. Real studies, often with real results, but which are partly based on unfounded elements”, regrets Guillaume Cabanac, whose tools made it possible to identify thousands of problematic articles.
These researchers hope that the verification tools they have developed will ultimately make it possible to clean up the scientific literature. In the meantime, they are increasing calls to publishing houses, still too few, according to them, to take stock of the problem. “Especially since it is not very complicated to carry out further verifications. The proof was enough for a well-informed specialist to read these studies,” argues Jennifer Byrne.
The specialist continues her research. Since the publication of her investigation, she has spotted other absurd mentions, suggesting that the fraud is extensive. What are these counterfeiters playing at? Hard to say. Unlike a judicial investigation, no police officer waits for crooked scientists to gather their version of the facts. But in research, the number of publications matters a lot in obtaining positions and funding. Copying articles and putting your signature on questionable transactions allows you to improve your statistics. Until some people took a closer look.
.