Automated segmentation and alignment of mitotic nuclei for kymograph visualisation

The mechanics of many biological processes can only be uncovered through the analysis of spatio-temporal data. Kymographs are a popular tool for visualising dynamic processes whose movements can be mapped into a single dimension, such as mitosis, or cell division. However, global movements of a cell means the region of interest (ROI) used to create the kymograph must move with each frame. Here we present an automated algorithm for the selection of dynamic ROIs to allow the construction of kymographs of proteins associated with the mitotic spindle in Drosophila embryos. We segment a dividing nucleus using a level set coupled to a shape model, which we then use to align the segmented nuclei to a common frame of reference.