Coordination of epithelial cells during morphogenesis

Epithelial cells are capable of sensing and reacting to the forces and movements of their neighbours. We fully quantify the dynamics of epithelial tissue using a novel high-throughput image analysis pipeline based on deep neural networks. Inspired by graph theory, we decompose cell-cell interactions into three distinct coupling types. With this approach, the epithelium can be represented by a planar graph of cell couplings whereby cells are interpreted as vertices and junctions between cells as edges. Utilizing graph-theoretic measures, we compare wild type tissue and mutants with impaired signal transduction capabilities revealing significant differences in e.g. composition of coupling types and spatial distributions. In contrast to the wild type, tension in those mutants is anisotropically distributed, indicating that local cell-cell coordination through mechano-sensing is essential for the function of an epithelium as force-generating tissue.