Making each individual positive pixel an object


#1

I used ImageJ to create a binary colocalization image. I want to be able to record each positive pixel as an individual object even if two positives are next to each other. I would like to do this because I am trying to get colocalization related to the perinuclear zone versus the cell periphery. Do you know how I could do this? Thanks


#2

I may not be understanding what you’re trying to caluclate, but rather than making single pixels objects, couldn’t you just identify objects based on the two signals that you’re analyzing for colocalization, identify the nucleus (I’m assuming you’re doing this already because you need to look at perinuclear colocalization), and as a bonus identify the cell (using identify secondary object (you only need to do this if you want to compare perinuclear versus cell periphery colocalization). Afterwards, you could use the expand module to expand out the nucleus that you identified. You could then use identify tertiary to create dis-union groups with your different regions (i.e. expanded nucleus -nucleus is perinuclear location). You could then use the dis-union groups to calculate colocalization of your signal withint those areas.

Sorry if this was a bit vague. Like, I said, I don’t totally know what you’re doing (or what you have to start with), so it’s hard to be too specific about things.

Cheers,

Peter


#3

I appreciate the advice. The problem is that I am looking for colocalization of a known protein pair in different cell types. The proteins do not form objects per se, so I can’t really identify them as objects. What I need is for each colocalization point a comparison of distance to nucleus and distance to cell periphery. I will then compare the regions of colocalization in the different cell types.


#4

Interesting! Okay, given you have an image where each positive pixel (=1) is a point of colocalization and each 0 pixels is an absence of colocalization, and you have a way to identify the cell and a way to identify the nucleus, why not use the measureradialdistribution module? Then you’ll be able to compare the fraction of colocalization close to the nucleus and the fraction close to the cell. Let me know if that helps at all.

Cheers,

Peter


#5

That is exactly what I needed, thank you so much. I had an older version of cell profiler that did not have that module. Thank you thank you thank you. That module will provide me with exactly the measurements that I need.


#6

Sweet! I’m glad that worked out :smile: