A paper posted to arXiv this week describes a new mathematical framework for tracking how cells move through tissue over time, not just where they end up.
The paper, arXiv:2609.20008, introduces a method called Travelling Pair Dynamical Alignment and Trajectory Estimation, or TP-DATE. It extends Gromov-Wasserstein optimal transport, a technique that already aligns static snapshots of tissue by weighing both gene-expression similarity and spatial structure. TP-DATE adds a dynamic version, estimating continuous trajectories between time points without running full physical simulations. On synthetic data and real spatial transcriptomics datasets, the paper reports that TP-DATE preserved tissue structure better and produced more accurate 3D reconstructions of cell dynamics than prior static approaches.
Spatial transcriptomics experiments are still snapshots: researchers slice tissue, measure gene activity, and lose the timeline in the process. A simulation-free way to fill in that timeline while respecting spatial structure could make it cheaper to study development, disease progression, or tissue repair over time.
It is a math paper, not a diagnosis tool; the real test will be whether biologists adopt frameworks like TP-DATE to interpret their own datasets, and that verdict is still out.