Sir

Your News Feature “CSI: cell biology” (Nature 434, 952–953; 2005) addressed an insidious, but largely ignored, problem with undocumented image enhancement in scientific papers.

The enthusiasm for presenting the ‘best’ scientific image possible seems to be driven by a desire both to tell a clear story and to compose an aesthetically pleasing image. It would be helpful if all journals adopted a code of image-manipulation ethics, such as those described in the News Feature for The Journal of Cell Biology, to guide authors and reviewers alike.

After teaching extramural microscopy and imaging courses for a number of years, I have observed two additional factors that contribute to the widespread manipulation of scientific images.

First, graduate school curricula typically do not offer systematic instruction in microscopy or image formation, with the result that most biology graduate students rely on ad-hoc training by more senior students or postdocs. Without comprehensive training, many junior scientists are unable to produce the quality of image desired and resort to image software manipulation to ‘fix’ the image. Developing expertise in image acquisition would be preferable to resorting to post-acquisition manipulation.

Second, your News Feature attributed the increase in questionable image manipulation practices to the eagerness of students and postdocs to improve their data.

However, in my courses, many trainees report that they are instructed — often pressured — by the principal investigator to produce images consistent with expectations. This often means losing the dynamic signal range inherent in biological material to create a high-contrast ‘unambiguous’ image. Such instructed manipulation, either in image acquisition or post-processing, essentially discards data at best and may be misleading at worst. Thus it is incumbent, not only on the scientist in training, but also on the scientist performing the training, to maintain high ethical standards while pursuing both beauty and truth.