Parameter for internal complexity/structure


In the idea to rank segmented objects with their internal complexity or structure, what would be the closest CP parameter that could reflect what, in flow cytometry, the SSC (side scatter) is? texture? granularity?




With regard to terminology, the closest is granularity. However, if you are using a “hybrid” instrument that really has SSC signal in images, such as Imaging flow cytometry, then imaging texture and radial intensity distribution would give you richer information about SSC signal than granularity.

If your goal is to compare a measurement in Flow cytometry and another one done by microscopy, I suggest to use both Granularity and Texture variance as a dual parameter. The reason is the cells that sit flatten on a microscopic slide have quite different SSC properties.



Thanks for your reply.

I’m “segmenting” blood cells, so I was wondering which CP parameter would be the closest to SSC to enable me to discriminate between lymphocytes, monocytes, neutrophiles, etc…like it is done by flow with FCS vs SSC dot plots.



One way to find out would be to measure a bunch of Granularity and Texture measurements, then do machine learning in CellProfiler Analyst to sort your cells into the correct categories- CPA would then tell you which measurements do the best job at discriminating these. :smile:

This tutorial provides an example of you can identify and analyze cells in CP and then easily export the data for use with CPA.


thanks, I will try that.