Colocalization

In fluorescence microscopy, colocalization refers to observation of the spatial overlap between two (or more) different fluorescent labels, each having a separate emission wavelength, to see if the different "targets" are located in the same area of the cell or very near to one another. The definition can be split into two different phenomena, co-occurrence, which refers to the presence of two (possibly unrelated) fluorophores in the same pixel, and correlation, a much more significant statistical relationship between the fluorophores indicative of a biological interaction. This technique is important to many cell biological and physiological studies during the demonstration of a relationship between pairs of bio-molecules.

The ability to demonstrate a correlation between a pair of bio-molecules was greatly enhanced by Erik Manders of the University of Amsterdam who introduced Pearson's correlation coefficient to microscopists, along with other coefficients of which the "overlap coefficients" M1 and M2 have proved to be the most popular and useful. The purpose of using coefficients is to characterize the degree of overlap between images, usually two channels in a multidimensional microscopy image recorded at different emission wavelengths. A popular approach was introduced by Sylvain Costes, who utilized Pearson's correlation coefficient as a tool for setting the thresholds required by M1 and M2 in an objective fashion. Costes approach makes the assumption that only positive correlations are of interest, and does not provide a useful measurement of PCC.

Although the use of coefficients can significantly improve the reliability of colocalization detection, it depends on the number of factors, including the conditions of how samples with fluorescence were prepared and how images with colocalization were acquired and processed. Studies should be conducted with great caution, and after careful background reading. Currently the field is dogged by confusion and a standardized approach is yet to be firmly established. Attempts to rectify this include re-examination and revision of some of the coefficients, application of a factor to correct for noise, "Replicate based noise corrected correlations for accurate measurements of colocalization". and the proposal of further protocols, which were thoroughly reviewed by Bolte and Cordelieres (2006). In addition, due to the tendency of fluorescence images to contain a certain amount of out-of-focus signal, and poisson shot and other noise, they usually require pre-processing prior to quantification. Careful image restoration by deconvolution removes noise and increases contrast in images, improving the quality of colocalization analysis results. Up to now, most frequently used methods to quantify colocalization calculate the statistical correlation of pixel intensities in two distinct microscopy channels. More recent studies have shown that this can lead to high correlation coefficients even for targets that are known to reside in different cellular compartments. A more robust quantification of colocalization can be achieved by combining digital object recognition, the calculation of the area overlap and combination with a pixel-intensity correlation value. This led to the concept of an object-corrected Pearson's correlation coefficient.

...
Wikipedia