Whole-cell particle-based digital twin simulations from 4D lattice light-sheet microscopy data.

We introduce a whole-cell digital twin framework that integrates four-dimensional (4D) (x, y, z, and t) lattice light-sheet microscopy with particle-based reaction-diffusion simulations in ReaDDy to model mesoscale intracellular organelle dynamics. Using fluorescence microscopy data from live Cal27 cells, we construct spatially resolved digital twins incorporating mitochondrial networks, microtubule networks, dynein and kinesin motors, the plasma membrane, and the nucleus. Mitochondrial dynamics include fusion/fission remodeling, diffusion, and motor-driven active transport along microtubules. Our simulations reproduce experimental trends in mitochondrial dynamics across control and two microtubule-perturbed conditions, demonstrating predictive capability without reparameterization. We then use stress-mimicking to predict emergent perinuclear mitochondrial clustering. Crucially, these simulations reveal that microtubule topology acts as a structural gate for this reorganization, demonstrating that upregulated retrograde motor kinetics alone are insufficient to drive clustering without permissive filament connectivity. This digital twin framework provides an approach for investigating intracellular dynamics and perturbation effects in an interpretable and biologically grounded manner.