Establishment of morphological atlas of Caenorhabditis elegans embryo with cellular resolution using deep-learning-based 4D segmentation

Cell lineage consists of cell division timing, cell migration and cell fate, which are highly reproducible during the development of some nematode species, including C. elegans. Due to the lack of high spatiotemporal resolution of imaging technique and reliable shape-reconstruction algorithm, cell morphology have not been systematically characterized in depth over development for any metazoan. This significantly inhibits the study of space-related problems in developmental biology, including cell segregation, cell-cell contact and cell shape change over development. Here we develop an automated pipeline, CShaper, to help address these issues. By quantifying morphological parameters of densely packed cells in developing C. elegans emrbyo through segmentation of fluorescene-labelled membrance, we generate a time-lapse framework of cellular shape and migration for C. elegans embryos from 4-to 350-cell stage, including a full migration trajectory, morphological dynamics of 226 cells and 877 reproducible cell-cell contacts. In combination with automated cell tracing, cell-fate associated cell shape change becomes within reach. Our work provides a quantitative resource for C. elegans early development, which is expected to facilitate the research such as signaling transduction and cell biology of division.