How to exclude background when too far from nuclei?


When analyzing the above picture the red/green represents cell type and a protein. The problem is that in the white circles (in the circular area with no nuclei)there are some background colors that are included as cell type and protein even if this should not be possible. It must be background.

My question is: Is there a way to limit the inclusion to red/green signal so that they have to be within XX pixels from a nuclei. The cells i’m looking at are T-cells which mainly are just nuclei with a cytoplasm very close to these.

Does the above explaination make any sense?

Best regards
Lars

I actually found a solution myself :smile:
If you think it’s wrong then please correct me. Just wanted to update if anybody else comes in the same situation as me:

I already did a “CD4 positive cells of all nuclei” (Filterobjects - Nuclei and CD4 Cells_count)
and a “KV1.3positive cells” (filter objects - CD4 cells and kv1.3channels_count)

when doing the Calculate math in the end the numinator should be the KV1.3 positive cells and the denominator CD4 positive cells (instead of just CD4 cells).

This solved the problem.

Yes, I think limiting your pipeline to just the pre-identified cells positive for the protein will do the trick.

Another approach is to find the nuclei, use the “Distance-N” method in IdentifySecondary to define a cell N pixels around each nucleus, identify the CD4+ regions and Relate them to the cells, and then use FilterObjects to find the cells which contain the CD4+ regions.
-Mark

Hi Mark (or anybody else if you know how) :smile:
I’ve tried to do as you suggested as this would be better. I did it my way but ran into trouble with background inclusions after all.

I’ve attached two pictures. One with the background i want to exclude and the other picture of some cells positive for both the RED(CD4) and GREEN(KV1.3) antibody.

I really need your help on this getting Cellprofiler to find the cells on the second image and excluding the background from pic 1.

Thanks in advance.




test.cp (12.2 KB)

Hi,

I’m attaching a pipeline that does what I described in my prior message. Also, I noticed that you are using a trunk build version of the pipeline; may I ask why, since there seems are no modules/features in the pipeline that is not accounted for by the current release?

Regards,
-Mark
2013_05_28.cp (11.5 KB)

Hi Mark.
I’m not really sure what you mean by trunk build. I’ve just created the pipeline in Cellprofiler?

Did you get your version of CellProfiler from this page? cellprofiler.org/cgi-bin/trunk_build.cgi
-Mark

I’m pretty sure that i just downloaded from the download page. Maybe i did an update from within the Cellprofiler software.

Should I uninstall and then reinstall the version from the download page.
My version is 2.0.0 rev 40aad83 (Mac)

That’s definitely not the current release; it should be version r11710. Give it another try with the Mac version from this page: cellprofiler.org/download.shtml.
-Mark