Cell tracking is an indispensable tool for studying development by time-lapse imaging. However, existing cell trackers cannot assign confidence to predicted tracks, which prohibits fully automated analysis without manual curation. We present a fundamental advance: an algorithm that combines neural networks with statistical physics to determine cell tracks with error probabilities for each step in the track. From these, we can obtain error probabilities for any tracking feature, from cell cycles to lineage trees, that function like P values in data interpretation. Our method, OrganoidTracker 2.0, greatly speeds up tracking analysis by limiting manual curation to rare low-confidence tracking steps. Importantly, it also enables fully automated analysis by retaining only high-confidence track segments, which we demonstrate by analyzing cell cycles and differentiation events at scale for thousands of cells in multiple intestinal organoids. Our approach brings cell dynamics-based organoid screening within reach and enables transparent reporting of cell-tracking results and associated scientific claims.

Fluorescence live-cell microscopy is one of the most frequently used techniques to study dynamic processes in organoids. However, it is often limited by laborious fluorescent reporter engineering, limited numbers of fluorescence channels, and adverse phototoxicity and protein overexpression effects. Label-free imaging is a promising alternative but not yet established for 3D cultures. Here, we introduce LabelFreeTracker, a label-free machine-learning-based method to visualize the nuclei and membranes in bright-field images of 3D mouse intestinal organoids. The approach uses U-Net neural networks trained on the bright-field transmitted light and fluorescence images of mouse intestinal organoids as obtained by standard confocal microscopy. LabelFreeTracker frees up fluorescence channels to study fluorescent reporters and allows (semi-)automated quantification of cell movement, cell shape and volume changes, proliferation, differentiation, and lineage trees. This method greatly simplifies live-cell imaging of tissue dynamics and will accelerate screening of patient-derived organoids, for which reporter engineering is not feasible.