Counting colocalized (or not) objects

Image Analysis
#1

Hi,
I’m very pleased to start using the CP. But as expected for a beginner i have had lots of challenges to put my pipeline to work.
The experiment:
I give some small bugs to cells to evaluate the phagocytosis.
The objective: simply know an infection index (cells with intracellular bugs)
The method: the bugs are GFP, the non-phagocytosed bugs are imunostained, the cells can be identified in the phase image. So i have 3 images per set.

I could identify all three objects independently, the problem is how to count per image the cells hosting phagocytosed bugs. To do that I need to count the cells that colocalize the only GFP bugs (not immunostained). The object for immunostaining are a bit bigger than GFP signal.

Please, could you help me.

#2

Provided that you are indeed able to identify the three sets of objects, to count the cells that contain GFP bugs, you can use the RelateObjects module to to establish a parent-child relationship between the bugs (children sub-objects) and the cells (parents). You can then use the FilterObjects module to filter cells by the children bug count, and retain those cells which have a count of 1 or more. In other words, those cells which have more than 1 child co-localize with at least 1 bug.

You can look at the cell/particle counting pipeline on our Examples page for an instance of how this works. Hope this helps!
-Mark

#3

Thank you. But there is still one problem: using this approach I’ll get the number of cells with Bugs(GFP), but there are two types of bugs: Stained(extracellular) and not stained (intracellular), how to count the cells that colocalize with bugs that are NOT stained (only GFP)?

#4

A similar approach will work here as well: Use RelateObjects to relate all the bugs (parents) and the stained extracellular bugs (children). Then use FilterObjects to remove the bugs which have 0 “children”, i.e., those that co-localize with no stain. This will work provided that the GFP image and the stain image are properly aligned so that GFP/stains co-localize correctly.

Once you have the GFP-only bugs, you can then use these objects in the approach I described previously.

Regard
-Mark

#5

Wonderful!

Thank you!