More pixels? not always better analysis!

Imaging in microscopy is fraught with places where a small change in digital images can make a big difference in results. Case in point, made simply with photoshop, is that having an image at 600 ppi doesn’t always give you more information than processing an image that is only 72ppi. It all depends upon the ratio of ppi to blur.

These three identical images, from a published article by Lu et al, on bovine surfactant protein D is what i am working on to determine what the differences in shadow cast images, AFM images, image processing and presence or absence of conserved peaks in the SP-D dodecamer structure.  Below, a very low res image from a publication was used to see what could be determined about size and structure using simple image processing (in this case Adobe photoshop).

The conclusion is that more pixels is not always better.  I think it is kind of cute that in this bovine SP_D the N terminal junctions of the dodecamer here are not quite like that of rhSP-D (shadowed images), and there is not a prominent dark area where one would like to see some indication of a glycosylation site. And also I find the three ends of the trimer carbohydrate recognition domains are very nicely shown right next to the three predictable peaks in the collagen like domain (as routinely seen with AFM images of rhSP-D) so thats just a nice treat.