A Fast Automatic Colocalization Method for 3D Live Cell and Super-Resolution Microscopy

Colocalizing two fluorescent-labeled proteins remains an open issue in diffraction-limited micro-scopy and raises new challenges with the emergence of super-resolution imaging, single molecule tagging (PALM, dSTORM...) and high content screening. Two distinct colocalization approaches are usually considered to address this problem : the intensity-based methods are very popular but are known to be sensitive to high intensity backgrounds and provide errors if the signal-to-noise ratio (SNR) is low ; the object-based methods analyze the spatial distribution of the two sets of detected spots by using point process statistics but unfortunately get rid of valuable information by reducing objects to points. We propose a unique method (GcoPS : Geo-coPositioning System) that reconciles intensity-based and object-based methods for various applications in both conventional diffraction-limited and super-resolution microscopy. Unlike previous methods, GcoPS is very fast, robust-to-noise and versatile since it efficiently handles 2D and 3D images, variable signal-to-noise ratios (SNR) and any kind of cell shapes and sizes. The experimental results demonstrate that GcoPS unequivocally outperforms the best competitive methods in adverse situations (noise, chromatic aberrations, ...). The method is able to automatically evaluate the colocalization between large regions and small dots and to detect significant negative colocalization. Since the one-parameter (p-value) GcoPS procedure is very fast in 2D and 3D, it should greatly facilitate objective analysis in large-scale high-content screening experiments