There’s fraud all along the chain of science, from the lab until final publishing. Tens of thousands of papers affected, the full extent unknown, but it’s widespread. Simple question: can any paper these days be trusted? Because even if the study researchers are 100% innocent, the fraud can happen upstream from them, starting with fake or adulterated lab supplies.
Lab supply companies have been selling antibodies using manipulated images
Basically, no.
And yes, fraud is rampant. I serve as an Associate Editor at a decent journal under a major publisher. Around 25% of manuscripts we receive are rejected due to failing basic automated checks for plagiarism (in this case, the automated checks only catch text re-use, not recycling of ideas/concept etc). I have personally caught at least 10 papers last year which were re-using data from previous publications, or publishing follow-on studies without citing the previous study - that’s like a falsification of novelty which interrupts the “chain of evidence” and replication idea.
I also just see several other things which are not plausible. n = 3 animals per group with tiny standard deviations on measurements. Real world doesn’t work like that. You can measure the ejection fraction %, blood pressure, serum hormone level etc several times in a row and not be as consistent as some papers claim 3 separate animals to be. So there is either fabrication, or some selective exclusion of data going on.
Overall it’s a massive problem. And AI is making things worse on both ends. Authors use it to more convincingly hide things, fabricate more plausible evidence etc, and the reviewers often outsource the review to AI models, which give “meh” reviews that end up with revisions.
The flip side is that AI tools have made it easier than ever to detect fraud. There’s a huge movement going on now to go through published papers and find problems. Somebody I know (and thankfully never collaborated with) has more than 20 papers with expressions of concern and a few retractions due to re-use of data across multiple studies spanning years.
And as you said, if the suppliers are lying to you about what is in the (very expensive) tube, then even honest researchers have no hope.