I am hoping that some useful information will come from the development of this method of assessing AFM images, in particular here, of SP-D. The measures are actually 8 in number but i am not showing but debating whether it is necessary.  All dodecamers have been normalized to 100 nm arm length. Just one image (not my AFM but found in the literature) is shown, a test of the methodology.

Below there are measures in square nm for the carbohydrate domains (shown in the top part of the image), values for the angle of separation of the arms (only one acute angle is shown here, mean of the acute and obtuse angles for this image are given below); peak width is given, and peak height (both in nm) (from the deepest part of the valley on one side), nm² for each peak (also from the deepest point on the peak), luminance is given (0-256) from the look up tables (obtained in ImageJ).  I understand that there are program s that can do some of these functions “automatically” but there are always details that machine vision really doesn’t “get right”.  This is an N of one, but this particular molecule was chosen (subjectively) as something very typical of AFM images purposefully in order to develop a technique.

So many things could easily be done…. for starters, comparing the nm2 of various CRD on the literally hundreds of CTLD (or even just the more well known c-type lectins, not to speak of arm arc angles and the position of the branching, and the N terminal connection width (in terms of nm).  All of the numbers below tell something about SP-D which need to be redone in dozens more images, and then compared to the variants of SP-D.