Colocalisation- need to quantify amount of protein on a particular cell type

Channel2 Channel1 Hi,

I have labelled astrocytes in the red channel and my protein of interest in the green channel. I would like to see whether the expression levels of the protein (green channel) expressed on astrocytes are different between groups.

I would therefore like to do a colocalisation analysis but I am not sure of the best way to do this? I am using imagej. This protein is expressed on other cells too, but my aim is to quantify the amount of the protein which is colocalised only on the astrocytes (so not to get a value of how much overlap there is).


There is a Coloc2 plugin in FIJI, but for anything more specific you would probably need to provide an image or clearer description. Having a DAPI channel, or some other basis for cell segmentation changes your options.

Thanks for the reply, I have attached images now (I also have a dapi channel). I have been reading about the coloc2 plugin, can this be used to quantify how much of the green is expressed on the red channel?


If you are looking for information like, how frequently is green in red, or how frequently is red in green, you can get those values from the Manders coefficient. You need both solidly meaningful thresholds and a good boundary.
I suggest reading the section on Fractional Overlap in this article:
The Coloc2 link above has some more specific information on using Manders coefficients and what can go wrong.

  • I would guess based on your image that you don’t have a channel that nicely outlines the total shape of the cell, so you may want to create a new channel as a sum (or something similar) of the red and green channels to find accurate enough outlines of your cells.