How to quantify spots from in situ?


#1

I would like to quantify colorectal cancer stained with DAB staining with RNAscope in situ. But I am not familiar with Cellprofiler, so I wonder if it is the right tool for me and which pipelines would be suitable for doing this. The staining looks like this: https://acdbio.com/rnascope®-20-hd-assay-brown-manual-assay

Thanks in advance!


#2

Hello Carl,

Those examples you linked looks “feasible” for CellProfiler. The general idea :

  • First use module “UmixColors” to extract the brown spots.
  • Then use “EnhanceOrSuppressFeatures” to enhance the appearance of the spots and suppress every other unwanted details
  • Then use “IdentifyPrimaryObjects” as usual.

You may find some hints in this example

Bests,
Minh


#3

Hello Minh!

Thank you for your reply! I have started experimenting and it seems promising. I have some issue since I would like to be able to choose spots with the highest intensity since there are some background that is counted as spots after unmix. This gets better if I initially use an image run through color deconvolution with Fiji.
Is it possible to adjust the intensity to only count the highest intensity spots?

Kind regards
Carl


#4

Hi,

It may be easier to first let it count all the spots. Then add MeasureObjectIntensity for the spot objects, then apply FilterObject in which you can freely choose a threshold of intensity to filter the unwanted spots.

This way you can easily re-visit and adjust cutoff threshold without repeating a whole process.

Hope that helps.


#5

Hi Carl,
You can also try using one of the Background/Robust Background threshold methods- they’re optimized for only picking out the very brightest areas of an image.