A novel method for quantified, superresolved, three-dimensional colocalisation of isotropic, fluorescent particles.

Colocalisation, the overlap of subcellular structures labelled with different colours, is a key step to characterise cellular phenotypes. We have developed a novel bioimage informatics approach for quantifying colocalisation of round, blob-like structures in two-colour, highly resolved, three-dimensional fluorescence microscopy datasets. First, the algorithm identifies isotropic fluorescent particles, of relative brightness compared to their immediate neighbourhood, in three dimensions and for each colour. The centroids of these spots are then determined, and each object in one location of a colour image is checked for a corresponding object in the other colour image. Three-dimensional distance maps between the centroids of differently coloured spots then display where and how closely they colocalise, while histograms allow to analyse all colocalisation distances. We use the method to reveal sparse colocalisation of different human leukocyte antigen receptors in choriocarcinoma cells. It can also be applied to other isotropic subcellular structures such as vesicles, aggresomes and chloroplasts. The simple, robust and fast approach yields superresolved, object-based colocalisation maps and provides a first indication of protein-protein interactions of fluorescent, isotropic particles.