This very thin and small peak is detected about 20% of the time on the downslope of the very large N term peak.  I checked to see whether the image processing peak finding was any different in terms of detection than the combined image and signal processing functions and they were virtually identical  (missing 79% of the time, detected 21% of the time).

Same layout and procedures as before. N term peak height data from 12 dodecamers of SP-D. Grayscale values for peak height and valley determined using ImageJ. N of 752 is the total number of trimers in 12 dodecamers using dozens of images, image filters, and signal processing functions.   INdividual dodecamers listed as well as a summar.