What intensity measurement is appropriated for my analysis?

I’m measuring fluorescence intensity using the cell profiler. However, my major problem is deciding if integrated intensity or median/mean intensity is appropriated.

Hi Mariana,

I don’t use CellProfiler, but I’d imagine that the difference is that total integrated intensity scales with area, while median/mean doesn’t. So if you want your intensity values to be larger for bigger objects, use integrated intensity; otherwise, mean/median. Hope that helps!


how can we know what is the appropriate intensity for your analysis if we don’t know anything about your experiment and your experiment goal?

please elaborate a little bit more…

1 Like

I’m using the cellprofiler to measure kinetics on activation of pkc using fluorescence (IF). Some cells seem to have an increase on fluorescence intensity, however it is not uniform for the whole population. In addition there are changes in sub-cellular localization.

Hi Mariana,
Sharing a picture could help us to answer.:wink:
Using integrated or mean will indeed affect your measurement. The cell spreading vs rounding, the depth of focus of your objective, the type of microscope (widefield vs confocal), uneven illumination are some of the parameters that could impact your measurement.
Furthermore, do you have an independent channel to detect the cells and then measure the intensity on your channel of interest? Otherwise this could also lead to introducing bias in you results.
Cf page



1 Like

Hi Romain,
Thank you for replying me. I use dapi to identify my primary object.

Hi Mariana,
So you detect pimary objects on the DAPI channel,
Then I guess you measure the intensities of the primary objects on another channel

You were mentionning before that :

If it’s about nuclear translocation, a common workflow is to :

  • expand the nucleus object by some pixels
  • create a tertiary object using the nucleus object and its expanded version (you get a ring around the nuclei).
  • Measure Mean intensity in the Nucleus VS the Ring

In that case, it could make more sense to measure “Mean Intensity” (because nuclei and ring could have very different areas).

Do the cells have different nuclei sizes ?
Here is another post about this topic

@CellProfilerTeam , is “integrated intensity” in cellProfiler equal to IntDen or RawIntDen in FIJI/ImageJ ?
IntDen = Mean * Area (hypothesis of a Gaussian Distribution)
RawIntDen = Sum Pixels Int (just a sum of all the pixels)

Again, sharing a picture would help us answering your question.



1 Like

@romainGuiet, in CellProfiler Integrated Intensity is sum of pixels’ intensities, i.e. equal to RawIntDen.

@mariana, in general chemical kinetics (including biochemical one) usually depends on the concentration of acting molecules. Most likely it would be so for you process as well.
Also, what kind of changes in sub-cellular localization are there? Is it translocation to nuclei, membrane, mitochondria, etc.?

1 Like