Cell morphological motif detector for high-resolution 3D microscopy images

Recent advances in light-sheet microscopy enable imaging of cell morphology and signaling with unprecedented detail. However, the analytical tools to systematically measure and visualize the intricate relations between cell morphodynamics, intracellular signaling, and cytoskeletal dynamics have been largely missing. Here, we introduce a set of computer vision and graphics methods to dissect molecular mechanisms underlying 3D cell morphogenesis and to test whether morphogenesis itself affects intracellular signaling. We demonstrate a machine learning based generic morphological motif detector that automatically finds lamellipodia, filopodia, and blebs on various cell types. Combining motif detection with molecular localization, we measure the differential association of PIP2 and KrasV12 with blebs. Both signals associate with bleb edges, as expected for membrane-localized proteins, but only PIP2 is enhanced on blebs. This suggests that local morphological cues differentially organize and activate sub-cellular signaling processes. Overall, our computational workflow enables the objective, automated analysis of the 3D coupling of morphodynamics with cytoskeletal dynamics and intracellular signaling.