Attempting to quantify tight junctions



Hello all,

Completely new to CellProfiler. I’m helping a colleague examine tight junctions from a fluorescent stain of claudin-5. Recently, I came across cell profile and saw a user posted around March 2011 with an example of the pipeline they generated. I tried to replicate a similar process, however to no avail. I simply can’t find a threshold value that doesn’t remove all intensity from the image (i.e. the image turns up black). Everything runs fine when I convert the image to grey and invert it. The next step is masking the image using the threshold- and that’s where I hit a snag. Any advice would be greatly appreciated.

Testpipeline.cpproj (219.0 KB)



  1. Could you upload the original image, not a jpg? (It’s fine if it doesn’t display in the browser window, and it’s also fine to upload it elsewhere and just link it here). Analyses tuned on one format might not work well on another format.

  2. Could you explain more clearly what you mean by this?

The next step is masking the image using the threshold- and that’s where I hit a snag.



So my colleague gave me a set of new files, in PNG format. I’m going to upload an example and hope that helps.

I was trying to find a way to isolate only the fluorescence of the tight junctions, but the rest of the image (that I originally uploaded) was also bright- so trying to attenuate the background also removed most of what I am trying to help my friend quantify (which was when I was saying thresholding the image has caused issues). Essentially, I am trying to find a way to quantify only the image intensity of the tight junctions, but I’m not sure where to start. I attached a PNG file for example.

EDIT: I attached a pipeline to kind of show what I’m trying to do, but it’s not very clean when I’m trying to detect tight junctions.

TJQuant.cpproj (639.3 KB)



I have made a couple of changes in your pipeline and perform a morphological operation (Closing) to enhance tight junctions followed by thresholding and masking image. I guess this works but you can fine tune pipeline as per wish. I have attached pipeline and the final image as you can see below. Let me know If this works for you.


TJQuant.cppipe (8.4 KB)