Computational Methods for Protein Localization Analysis

Fluorescence microscopy has enabled imaging of spatial proteome (morphological pattern of subcellular protein localization). Automated (and even manual) high-resolution fluorescence image acquisition generates a large amount of complex image data, and manual analysis of such data in order to distill biologically meaningful information is challenging. Automated image analysis is an inevitable approach to perform quantification of phenotypic changes, evade subjective bias, and provide accurate and reproducible results. Automation requires the utilization of image processing, image analysis, and data analysis tools. In Chap. 9 is presented our customized system for automated analysis of fluorescently stained blood cells. Our aim was to develop an automatic image analysis system that would enable us to minimize the amount of required manual operations not only throughout the inspection of segmentation results but also during the initial tuning of various parameters of the segmentation algorithm. Setting free from the tuning of initial parameters allows for faster switching to data analysis of the new experiment when imaging settings or other conditions were changed. The development of the fluorescence image analysis algorithms included a good practice from the development of 2DE image analysis system (see Chap. 6). The developed tools were applied for automated analysis of confocal microscopy images to evaluate changes of histone modifications in cell populations.