It is hard to reproduce the values returned by various instruments. Using modern technology and analytics tools it’s easier to check for fraud in the western blots, so this is where the fraud is often proven. The problem is that this is ipso facto fraud - you can’t say “ohhh I got mixed up and accidentally made an image that perfectly lined up with my conclusions”. If the authors were willing to fraud there then the entire paper should be seen as fraud and the authors responsible shamed because they are undermining the fundamental scientific process.