Automated analysis of secretory vesicle distribution at the ultrastructural level

Abstract

Neuroendocrine cells like chromaffin cells and PC-12 cells are established models for transport, docking and secretion of secretory vesicles. In micrographs, these vesicles are recognized by their electron dense core. The analysis of secretory vesicle distribution is usually performed manually, which is labour-intensive and subject to human bias and error.We have developed an algorithm to analyze secretory vesicle distribution and docking in electron micrographs. Our algorithm automatically detects the vesicles and calculates their distance to the plasma membrane on basis of the pixel coordinates, ensuring that all vesicles are counted and the shortest distance is measured. We validated the algorithm on a several preparations of endocrine cells. The algorithm was highly accurate in recognizing secretory vesicles and calculating their distribution including vesicle-docking analysis. Furthermore, the algorithm enabled the extraction of parameters that cannot be measured manually like vesicle clustering.Taking together, the algorithm facilitates and expands the unbiased and efficient analysis of secretory vesicle distribution and docking.