Help with Fiji Coloc2 and thresholding

Hi all,

I’m using coloc2 for the first time and I’m a bit confused RE: thresholding. I am analysing z stacks of 2 proteins that we see do not particularly colocalise (they tend to localise just because they are so spread out rather than being specifically localised). As they are not individual events in the cell - both proteins are spread out - one primarily in the nucleolus, the other in the cytoplasm - there is no specific ROI.

I import the images, split the channels to get red/green and create maximum intensity projections (my images are deconvolved). However, since I have no specific ROI as the proteins are diffuse, how do I deal with background and in what way should I be thresholding? I have read up on the colocalisation threshold but am unsure how to implement this in the actual analysis. Similarly, I am aware of background subtraction but am still unsure which is best to use. I am guessing I need to deal with the background but also know Pearson’s (which I am seeking to use) does not normally require thresholding.

Thank you!


I’m so sorry for the late reply! I must have overlooked this post… my apologies.

I’m worried that analyzing a z-projection may result in misleading data… as you can imagine - pixels from very different slices (ie - not near each other at all) - might be calculated as having ‘colocalization’. Would it be possible to either 1) select a representative slice (at the center of the cells?) or 2) analyze slice-by-slice?

To better help you - please attach/share an ORIGINAL image file with us here by uploading directly on the forum here or via a link to a file-sharing site (such as Dropbox).

Having a look at your data will help us better help you.