Colocalization studio: radius of colocalization

[Alanah Pieters]

Hello,

I'm currently using colocalization studio, but I'm having a problem with the diameter of colocalization.
I need to find out if my spots of colour 1 colocalize with my spots of colour 2. When I go to 'colocalization studio', I chose 'method: object'. Here, I have to give in 'maximal distance for the analysis in pxs'. (I suppose that this diameter I chose starts in the center of the spot of colour 1?). The problem lays in the fact that, when I take a smal radius, for exemple 5 pxs, I have more colocalization then when I take a larger radius, for exemple 6 pixels. This is strange, because normally, the larger the diameter, the more spots you will detect in this perimeter, so, the more colocalization... Can someone help me resolve this?

Greetings,

Alanah

[Thibault Lagache]

Hi Alanah,

Yes, you should have more colocalization (and actually reach a plateau at some distance, meaning that you have no more coloc).

However, the ethod implemented in coloc studio and described in Lagache et al. Cytometry Part A 2015 uses model fitting of Ripley's curve for estimating the coloc. There is an option in the plugin to check (by eye..) that the fitting curve is "close" to the measured Ripley's K function.

You may first check that fitting is ok in the 2 conditions

Then, If both are correct, subtle differences can appear because curve fitting remains a bit unpredictable. This means that if you reach the coloc plateau at 5 pixels, then it can slightly decrease for higher distances due to fitting artefacts

All the best

Thibault Lagache