Pipeline for fluo quantification from 2 fluo colors images


Lovely software, thanks a lot.

May I check with you if I’m correct in my thinking?

I have microscopy images of lymphocytes stained for 2 markers, therefore 2 fluorescences: green and red. The red fluo identifies the cell, the green is the marker of interest. I would like to quantitate the positive cells and quantify the green fluorescence within each red object, basically doing the job of a flow cytometer.
What would be the pipeline for that? Do I need to separate each color (ColorToGray)? do any illumination correction? IdentifyPrimaryObjects? IdentifySecondaryObjects? MeasureObjectIntensity? ExportToSpreadsheet?

Many thanks for your help.

It seems the basic steps in a pipeline like this are:

  1. import images
  2. Convert images to grayscale
  3. Correct illumination calculate and apply
  4. Identify primary objects, here they seem to advise identifying the nuclei first
  5. Identify secondary objects based on the objects from primary objects, these would be the cells
  6. run measure object intensity on object of interest
  7. run calculate math for ratios etc…
  8. export results to spread sheet.
1 Like

Awesome, @mplace!

I’d only like to add that if @vblanche does not have a DNA stain, then you will just run step 4 on the red channel and skip step 5.

Thank you mplace, thank you Anne.

I was more or less going towards the suggested pipeline. I’m still struggling with the illumination correction, I have to do it for the green channel (alexa 488) but not the red one (celltrace red).
I’m correct to think that I can use “IdentifyPrimaryObjects” on the red channel and use the “red” images created by the “ColorToGray” module, and then ask for the green fluorescence intensity of each “red” object? how I do that? When I use the “MeasureOjectIntensity” module, should I select the corrected green image and select the objects defined by the “identifyPrimaryOjects” module?

So, my pipeline could be something like (I’m using CellProfiler 2.2.0) (please, I will greatly appreciate any correction):

  1. ColorToGray
    create 2 grayscales images for both fluo channels: green and red

  2. CorrectIlluminationCalculate
    background, block size 20, no rescale, no smoothing

  3. CorrectIlluminationApply
    apply on green image, substract

  4. IdentifyPrimaryObjects
    create objects called “cell” from the red image; here I’m still playing with the threshold strategy and method to distinguish clumpled objects.

  5. MeasureObjectIntensity
    on corrected green image and selecting objects “cell”

  6. MeasureOjectSizeShape

  7. DisplayDensityPlot

  8. OverlayOutlines

  9. DisplayDataOnImages
    showing integrated intensity on original color image with outlines of cell objects.

  10. SaveImages

  11. ExportToSpreadsheet

What would be the closest measurement I could plot for size (equivalent of FSC in flow cytometry)? AreaShape_Area?
Can I plot Object diameter against IntegratedIntensity?

Many thanks,


Could I upload here my current pipeline and have your comments on it?

In fact, I’m kind of trying to do some colocalization of 2 fluo markers in lymphocytes, so I’m testing the pipeline from the ExampleColocalization tutorial. My question: can I use my color image and do a ColorToGray split before?

Many thanks

Sure, post your images (if they are too big, put them somewhere accessible on the web) and .cpproj file. I won’t likely have time to look myself but someone else might!

many thanks.

Here are an image and a pipeline:

test.cppipe (13.4 KB)