Sir

I was interested to see your News Feature “CSI: cell biology” (Nature 434, 952–953; 2005) on digital photography and image manipulation in cell biology. Photoshop-based enhancement of images raises questions of proper conduct, for which journal guidelines are necessary.

In my field, the problem (and it is here a problem, not an issue of misconduct) is much greater in electron than in light microscopy.

An example of good practice is the cover of the 2 December issue of Nature. This enhanced image is taken from an Article by A. Fotin and colleagues on clathrin lattices (Nature 432, 573–579; 2004). The authors are to be congratulated for their unusually open disclosure of the difference between the original and the published image.

The Methods section of this Article makes a clear distinction between the data collected by the cryo-microscope and the pictures in the article. The authors list all the image-processing programs used in preparation of the illustrations: IMAGIC, CTFTILT, FREALIGN, O, EMAN, Chimera, MAVE, LSQ_EXPLICIT, MAMA and MODELLER. These are mostly ‘off the shelf’ programs that produce symmetry enhancements, density averaging and many of the same effects as Photoshop.

In my experience, unless the scientist/postdoc/technician knows a great deal more about the guts of these programs than most, they are performing ‘black-box’ image enhancements that they do not control to any significant degree.

The full disclosure by Fotin and colleagues is remarkable for being so rare. The scarcity of such imaging disclosures elsewhere in the published record shows us just how far we have come towards inverting the purpose of scientific images. Where we used to have “seeing is believing”, we now have the possibility of “believing is seeing”, courtesy of our image-processing and enhancement software.