SKM 2023 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
BP: Fachverband Biologische Physik
BP 4: Tissue Mechanics I
BP 4.7: Talk
Monday, March 27, 2023, 16:45–17:00, BAR Schö
Tracking and comprehending single cell dynamics in Drosophila dorsal closure using machine learning — •Daniel Härtter1,3, Yuxi Long2, Janice Crawford2, Daniel P. Kiehart2, and Christoph F. Schmidt1 — 1Department of Physics and Soft Matter Center, Duke University, USA — 2Department of Biology, Duke University, USA — 3Department of Pharmacology and Toxicology, Göttingen University Medical Center
Dorsal closure in Drosophila melanogaster embryos is a key model system for cell sheet morphogenesis and wound healing. We pursue a data-driven approach to understand the emergence of organized behavior on tissue level from the stochastic dynamics of single cells across scales. We developed DeepTissue, a deep-learning-based algorithm to automatically and robustly detect and temporally track various single cell features: cell shapes, cell junction lengths, myosin intensities, and tissue topology. Epithelial cells in dorsal closure exhibit oscillations and contribute to progressive cell sheet movements, while showing a large variability in individual shapes, dynamics, and fates. Based on high-quality multi-parametric trajectories of 1000s of single cells, we use unsupervised machine learning techniques to detect and classify behavioral and structural phenotypes. Further we study how the behavior of single cells throughout closure is driven by deterministic and/or stochastic factors, with the aim to predict singular cell ingression events.