Colocalization Colormap
Independent fluorescent labeling of two different components is a commonly used method to determine their colocalization. This website contains an ImageJ plugin for automatic quantification and visualization of colocalized fluorescent signals.The plugin is based on the method, which was originally described by Jaskolski et. al. (2005). This algorithm computes correlation of intensities between pairs of individual pixels in two different channels. Results are presented as a colocalization color map. Hot colors represent positive correlation (colocalization), whereas cold colors represent negative correlation (exclusion).
Spatial representation of colocalization
Correlation between pair of pixels (described by x,y coordinates in the image) is calculated according to the formula presented below:  Distribution of nMDP (normalized mean deviation product) values (ranging from -1 to 1) is visualized with color scale. Indexes below 0 are represented by cold colors (exclusion). Indexes above 0 are represented by hot colors (colocalization). |  Fig.1 Automatic colocalization calculation using Colocalization Colormap plugin.
(A) Neuronal body (asterisks) and dendritic processes (arrowheads) stained with antibody against MAP2 protein.
(B) Immunoreactivity of CD44 adhesion molecule in the brain.
(C-D) Color map demonstrating colocalization between adhesion molecule and MAP2 neuronal marker. |
Statistical evaluation
Basic nMDP statistics was supplemented to the Colocalization Colormap plugin.
Statistical analysis of nMDP distribution might be used to compare the degree of colocalization between different experimental situations related to changes in fluorescent signal localization. The method is sensitive enough to detect even small differences in fluorescence distribution (see simulated data below).
Plugin calculates: Icorr (correlation index -fraction of positively correlated pixels in the image), the number of positive nMDP values, average positive nMDP value ( an arithmetic mean of positive nMDP data set), variance of positive nMDP values (how far the positive nMDPs lie from their arithmetic mean) and nMDP histogram data.
Simulated data
Method implemented to the
plugin was tested on artificial images generated to simulate gradual and progressive separation of two
fluorescent labels (as shown below in Fig. 2). To do this, confocal
image was duplicated and then the duplicated image was rotated 2º,
4º, or 6º
clockwise. Maximum projections of this two artificial channels were
than analysed with Colocalization Colormap plugin to create maps of
colocalization and calculate corresponding nMDPs distribution.
Fig. 2 Statistical evaluation of colocalization using Colocalization Colormap.
A-P) Pixels in original (red channel) confocal images were translocated (green channel) to simulate progressive separation of overlapping fluorescent labels. Q-T) Color maps demonstrating gradual decrease of colocalization (hot colors) between analyzed objects.
U) nMDP distribution in simulated images demonstrating the most (complete -red) and the least (partial 6° -blue) significant colocalization.