As of this minute, i am thinking that the number of peaks along the arms of the multimer under current investigation is along these lines.  One, eight, four.  That is, one (1) very bright peak (junction at the ring of all the multimer arms yielding one very bright peak), eight (5-8) regularly spaced pretty evenly sized less bright peaks (along the straighter part of the arm) and four (4) varying peaks, with one or two of the peaks being brighter than the other two, and typically the last peak being of lesser brightness than the three prior peaks, sort of a tapering off).  This has become apparent through “staring” at this molecule for several months, and processing the images in as many ways as made sense.  The easy way to count the peaks is by using this set of algorithms…  three different programs, sometimes four (photoshop, gwyddion, coreldraw, imageJ).  see image above (cropped_300ppi_psd_high_pass_250_gaussian_blur_5_px_0ther_maximum_10px-radius) and below (2Dfast fourier transform in gwyddion).  The plots are different but peaks look similar, the top image much easier to count peaks and a plot with less noise.

This is a consensus graph (plot) made from the summary of about a dozen different segmented lines overlying peaks in the outer arm of a multimer, which i am calling the “hook” portion.  This hook is a consistent and interesting feature and it occurred to me that there might be a pattern in the position of the brightest peak in these last four peaks.  There still may be, but it is not as definite as i had hoped.  This is, keep in mind, a summary of arms from just one multimer.  plots peaks are approximate, but the trend in rising peak heights is consistent.  The total arm length including this hook portion is something around 100nm, but yet to be confirmed. The description of how the images was processed is found here.

The purpose of singling out the hook was to try to find a common point from which to measure the straighter portion of the multimer arms and determine whether there is a repeating periodicity to the brightness peaks there.  It appears that a point just AFTER the four peaks in the hook might provide a good place to begin that search. It is disappointing not to have found a consistently, brightest peak.

An image provided to me to “look over” is in fact a wonderful gift (non-tangible, unexplainable, artistic, important).  I have spent a couple months trying to decipher what this protein actually might have as a consistent shape in its numerous arms. So i decided to look at one place where I thought i could tell there was a pattern in brightness peaks, and that was the “hook” at the end of each multimer.

image below has had the “hooks” cropped out of the original image (which stands as i have named it –round_1_orig_cropped_300ppi_psd_high_pass_250_gaussian_blur_5_px_limit_range_100-255 which means that it is the original image, increased to 300ppi in photoshop, and then to it applied a 250 high pass filter, then a gaussian blur of 5px, saved, then opened in gwyddion and in the latter, a limit ranger algorithm for basic operations>limit range (100-2555) applied. Images were cut and rotated in corelDRAW, then exported as a single image straight line, each identified from right to left with their tracings and their plots. ImageJ was used to create the tracings (segmented line), brightness peaks (plot profile). Plots saved to excel, imported as metafiles in CorelDRAW, and the brightness peaks kept relative to 100% and the x axes as nm according to the original bar marker. Bar marker was retraced with each segmented line of a hook area.

Mean peak width (+SD) in the area plotted in the red box with the plot line in red  (n=14, 36.5nm + 4.39); mean number of peaks in a hook (n=14, 4.2 + .69); right to left, the width of peak 1 (n=14, 9.5nm + 2.7); width of peak 2 (n=14, 7.28n + 2.4); width of peak 3 (n=14, 8.85 + 2.4); width of peak 4 (n=13, 8.30nm + 2.7); peak 5  (n=5, 7.2nm + 1.9);  It is possible to begin at the most right part of the tracing and count highest peak in that direction rather from left to right…. go right to left.  All tracings were done in the same direction from a portion of the molecule’s arm to the end of the hook. It is clear that a few of these have some ambiguity in where the hook hooks up to the arm…. that was a judgement call – in particular hooks labeled 1, 3 -5.  3 and 4 were traced twice.  Whether one bright peak was actually two spots was also a matter of judgement occasionally, but a 5% change in brightness was used to determine whether a peak was significant.
If i take out the duplicate of plots 3 and 4, and delete odd plot 1, then the picture is much clearer. peak number (n=11, 4.09 + 0.5); peak total hook width (n=11, 34.65 + 3.8); brightest peak  is either peak 1 or 3 (occurrence is 4 each out of n=11); peak 1 (n=11, 9.72 + 2.8); peak 2 (n=11, 7.8 +2.3); peak 3 (n=11, 9.0 + 2.44); peak 4 (n=10, 8 + 2.93); peak 5 (n=2, 8.5 + 2.5).  It looks like the brightest peak might be a little wider (obvious) and the the hook is 4 peaks.