Other papers that Dr. David found to contain irregularities were based on data generated in laboratories other than those of the Dana-Farber scientists, Dr. Rollins said. He said the institute had begun reviewing possible data errors in some of the cases pointed out by Dr. David even before he published a blog post about them on Jan. 2, or the Harvard Crimson published a story several days later. He also said that a review of three of the manuscripts highlighted by Dr. David did not support allegations of data irregularities.
“The presence of discrepancies in the images in an article is not evidence of the author's intent to deceive,” he added. “That conclusion can only be drawn after careful and fact-based examination, which is an integral part of our response.”
Dr David, who earned a PhD in cellular and molecular biology from the University of Newcastle in Britain, said he routinely studied scientific papers in his spare time. Despite having a single low-cost supporter on the online platform Patreon (which represents the only income he earns from his work), he estimated that he had left around 2,000 comments on PubPeer, a website where scientists provide public comments on the studies.
While he often digs into low-quality research conducted in China, Dr. David said, he had recently been investigating researchers collaborating with American scientists whose studies he had previously found flawed. After stumbling upon irregularities in the work of some Dana-Farber scientists, he visited the institute's website, began scrolling down its leadership page, and then dove into those researchers' papers.
He identified some problems with the help of artificial intelligence software before checking them himself. He found others alone. Other PubPeer users had pointed out some of the problems years earlier.
“There is a group of people in an institution with image problems,” Dr. David said. “How many mistakes are we happy for people to make and just say, 'That's an innocent mistake'?”
Among the most common irregularities he found were results from a Western blot, a method used to detect proteins, that were copied and pasted across different experiments, wrongly suggesting that a particular protein had been identified. In other cases, he said, images appeared copied, rotated or stretched in a way that suggested deliberate manipulation.
In one study, he found that an image of mice from the first day of an experiment appeared to reappear in the results on day 16, in a different part of the experiment.
“I'm concerned about the entire process, more than specifically this or that transfer,” Dr. David said. “Clearly something went wrong along the way.”
The findings also highlighted conflicts of interest that can complicate internal reviews of data irregularities. Dr. Rollins, Dana-Farber's Research Integrity Officer, is the author of some of the articles noted by Dr. David. The institute said he would be excluded from any investigation related to his research.
In the case of Dr. Glimcher, the institute's executive director, a committee of administrators will make a final decision on any findings.
