During embryogenesis tissue layers undergo morphogenetic flow rearranging and folding into specific shapes. While developmental biology has identified key genes and local cellular processes, global coordination of tissue remodeling at the organ scale remains unclear. Here, we combine in toto light-sheet microscopy of the Drosophila embryo with quantitative analysis and physical modeling to relate cellular flow with the patterns of force generation during the gastrulation process. We find that the complex spatio-temporal flow pattern can be predicted from the measured meso-scale myosin density and anisotropy using a simple, effective viscous model of the tissue, achieving close to 90% accuracy with one time dependent and two constant parameters. Our analysis uncovers the importance of a) spatial modulation of myosin distribution on the scale of the embryo and b) the non-locality of its effect due to mechanical interaction of cells, demonstrating the need for the global perspective in the study of morphogenetic flow.

The actomyosin cytoskeleton is a crucial driver of morphogenesis. Yet how the behavior of largescale cytoskeletal patterns in deforming tissues emerges from the interplay of geometry, genetics, and mechanics remains incompletely understood. Convergent extension in D. melanogaster embryos provides the opportunity to establish a quantitative understanding of the dynamics of anisotropic non-muscle myosin II. Cell-scale analysis of protein localization in fixed embryos suggests that gene expression patterns govern myosin anisotropy via complex rules. However, technical limitations have impeded quantitative and dynamic studies of this process at the whole embryo level, leaving the role of geometry open. Here we combine in toto live imaging with quantitative analysis of molecular dynamics to characterize the distribution of myosin anisotropy and the corresponding genetic patterning. We found pair rule gene expression continuously deformed, flowing with the tissue frame. In contrast, myosin anisotropy orientation remained approximately static, and was only weakly deflected from the stationary dorsal-ventral axis of the embryo. We propose that myosin is recruited by a geometrically defined static source, potentially related to the embryoscale epithelial tension, and account for transient deflections by cytoskeletal turnover and junction reorientation by flow. With only one parameter, this model quantitatively accounts for the time course of myosin anisotropy orientation in wild-type, twist, and even-skipped embryos as well as embryos with perturbed egg geometry. Geometric patterning of the cytoskeleton suggests a simple physical strategy to ensure a robust flow and formation of shape.

During morphogenesis, diverse cell scale and tissue scale processes couple to dynamically sculpt organs. In this coupling, genetic expression patterns and biochemical signals regulate and respond to mechanical deformations to ensure reproducible and robust changes in tissue geometry. A long standing approach to characterize these interactions has been the construction of expression atlases, and these atlases have necessarily relied on fixed snapshots of embryogenesis. Addressing how expression profiles relate to tissue dynamics, however, requires a scheme for spatiotemporal registration across different classes of data that incorporates both live samples and fixed datasets. Here, we construct a morphodynamic atlas that unifies fixed and live datasets, from gene expression profiles to cytoskeletal components, into a single, morphological consensus timeline. This resource and our computational approach to global alignment facilitate hypothesis testing using quantitative comparison of data both within and across ensembles, with resolution in both space and time to relate genes to tissue rearrangement, cell behaviors, and out-of-plane motion. Examination of embryo kinematics reveals stages in which tissue flow patterns are quasi-stationary, arranged as a sequence of morphodynamic modules. Temperature perturbations tune the duration of one such module, during body axis elongation, according to a simple, parameter-free scaling in which the total integrated tissue deformation is achieved at a temperature dependent rate. By extending our approach to visceral organ formation during later stages of embryogenesis, we highlight how morphodynamic atlases can incorporate complex shapes deforming in 3D. In this context, morphodynamic modules are reflected in some, but not all, measures of tissue motion. Our approach and the resulting atlas opens up the ability to quantitatively test hypotheses with resolution in both space and time, relating genes to tissue rearrangement, cell behaviors, and organ motion.

The actomyosin cytoskeleton is a crucial driver of morphogenesis. Yet how the behavior of large scale cytoskeletal patterns in deforming tissues emerges from the interplay of geometry, genetics, and mechanics remains incompletely understood. Convergent extension flow in D. melanogaster embryos provides the opportunity to establish a quantitative understanding of the dynamics of anisotropic non-muscle myosin II. Cell-scale analysis of protein localization in fixed embryos suggests that there are complex rules governing how the control of myosin anisotropy is regulated by gene expression patterns. However, technical limitations have impeded quantitative and dynamic studies of this process at the whole embryo level, leaving the role of geometry open. Here we combine in toto live imaging with quantitative analysis of molecular dynamics to characterize the distribution of myosin anisotropy and corresponding genetic patterning. We found pair rule gene expression continuously deformed, flowing with the tissue frame. In contrast, myosin anisotropy orientation remained nearly static, aligned with the stationary dorsal-ventral axis of the embryo. We propose myosin recruitment by a geometrically defined static source, potentially related to the embryo-scale epithelial tension, and account for transient deflections by the interplay of cytoskeletal turnover with junction reorientation by flow. With only one parameter, this model quantitatively accounts for the time course of myosin anisotropy orientation in wild-type, twist, and even-skipped embryos as well as embryos with perturbed egg geometry. Geometric patterning of the cytoskeleton suggests a simple physical strategy to ensure a robust flow and formation of shape.