Over the past ten years, a number of automated systems for acquisition and analysis of fluorescence microscope images have been described. These efforts have been primarily directed towards image cytometry [1,2,3], in which the goal is obtaining accurate measurements of the total fluorescence of each cell, or towards automating fluorescence in situ hybridization [4,5], in which the goal is determining the number of fluorescence "spots" (chromosomes) in each cell. While some image cytometry systems provide the ability to calculate numerical features from the fluorescence distribution for each cell, these are usually used to identify cell types (e.g., distinguish lymphoid from myeloid cells) [1,5] rather than to describe the subcellular pattern per se. Thus, features appropriate for describing protein localization patterns have not been previously characterized in the context of fluorescence microscopy.
An orthogonal approach to protein localization is not to describe it, but to predict it based on amino acid sequence. Work in this area has shown limited promise and there are currently no methods adequate for making specific predictions about the localization of a particular protein. One system was only able to predict localization correctly for 60% of the proteins studied [6]. Furthermore, that system and another [7] only placed proteins into general, non-specific categories (intracellular, extracellular, membrane associated, etc.). At this time, prediction of protein localization is not a viable alternative to experimentally determining and numerically describing the localization patterns themselves.