UBC Math Bio Seminar: Wenzheng Shi

Cells constantly reorganize the actin cytoskeleton to break symmetry, align with neighbors, and push their edges forward. There is a lot of quantitative data on these processes, but turning such data into testable mechanistic models remains a challenge.

I will first focus on collective cell chirality. Experiments show that single-cell actin chirality scales up to a tissue-wide chiral pattern on adhesive islands. Starting from observations of single cells, we formulate two different models: (i) inner cells rotate chirally, and chirality propagates outward; (ii) boundary cells tilt chirally at the edge, and chirality propagates inward. We simulate the models to propose experiments in which the two mechanisms predict opposite scaling trends. Experiments reveal that chirality first emerges at the boundary and propagates inward, selecting the “outside-in” model.

I will then turn to the leading edge protrusion of migrating cells. Multiplex microscopy provides local time series for concentrations of two key proteins - actin and Arp2/3 - together with edge velocity. We first treat these data in a model-agnostic way, using linear regression and phase-space analysis to infer an effective stochastic ODE system describing coupled actin–Arp2/3–velocity dynamics. We then propose several nonlinear mechanistic PDE models based on this inferred linear dynamics, prior biological knowledge, and parsimony. This class of models not only reproduces known features of the cell's leading edge but also explains the surprisingly high level of molecular noise.

These two studies showcase a common workflow for mathematical cell biology: infer simple mechanistic models from initial data, predict experiments allowing to discriminate between the models.