Help needed in tissue analysis pipeline?




I am trying to analyze tissue images with CellProfiler version 3.1.5.
I stained tissues with DAPI (nucleus), SIV envelope (TRITC), CD3 (FITC) and CCR6 (Cy5).
My goal is to identify all the infected cells (i.e., positive in TRITC) and check whether they are expressing CD3/CCR6 or both.
I am able to segment nucleus but I am not sure how to proceed further with other three channels.
I tried to use median or mean intensity of TRITC channel to identify the infected cells and it worked but when I analyze other images it doesn’t work.
How can I overcome the intensity heterogeneity in channels so that my pipeline works with other images as well?
I am uploading pipeline and corresponding image file for reference.
Pipeline_s4_2018.10.24.cpproj (1.7 MB) (1.1 MB)

Waiting for suggestions.


Rather than filtering objects using hard cutoffs in the FilterObjects modules, try using the Threshold module to identify TRITC-positive areas of the image (which will take into account the heterogeneity of various fields) and then use MaskObjects to select only objects that overlap with the TRITC-positive areas. I often find that is a much more robust strategy!


Thanks Beth for your suggestion.