bioRxiv2023
Microscopy has rapidly evolved at pace with live markers, enabling ever higher spatiotemporal resolution of multicellular dynamics within larger fields of view. Consequently, we are now in the era of widespread production of terabyte (TB)-sized time-lapse movies of experimental model systems, including developing embryos and organoids. Working with these large datasets has presented a new set of hurdles, particularly due to the lack of standardized open-source pipelines for acquiring, handling and analyzing the data. Moreover, although long-term tracking of a cell throughout an entire process, for example vertebrate organogenesis, is key to revealing the underlying cellular dynamics, this has proven largely elusive. To specifically address the question "But, what are the cells doing?", we created an image analysis pipeline optimized to track single cells in light-sheet acquired datasets (1 TB sized time-lapse, 8h of imaging, 30 min gene expression cycle, cell movement speed (1{micro}m /1 minute), 200-400 {micro}m tissue depth). Our modular pipeline optimizes and connects the following: image acquisition parameters to improve tracking feasibility; hardware specifications; data handling and compression tools; pre-processing steps; state-of-the-art cell tracking tools (Mastodon, MaMuT) and a novel open-source/ python-based tool (Paleontologist) to analyze and visualize spatiotemporal dynamics of the tracked cells. Importantly, our pipeline is adaptable to a variety of experimental systems and accessible to researchers regardless of expertise in coding and image analysis. One-sentence SummaryUser-friendly cell-tracking pipeline that connects image acquisition in multicellular systems through to data analysis of cellular dynamics.