UBC Math Bio Seminar: Felix Zhou

Form is function. Just as Darwin’s finches have beaks adapted to their ecological niche, so too has cell morphology adapted to the cell’s function and its local microenvironment. Consequently, cell shape changes are prominently used as phenotypic biomarkers of disease and an aggregate downstream readout of a cell’s molecular processes. However, high-resolution microscopy combined with advanced image analysis increasingly reveal a more complex interplay whereby cell shape plays a pivotal role in spatially organizing, and translating molecular signals to cell behavior. For example, we recently discovered a previously undescribed role of blebbing (local, dynamic, hemispherical surface protrusions) in melanoma cells to promote membrane recruitment of curvature-sensitive scaffolding Septin polymers, to amplify pro-survival signaling, enabling the cells to bypass anoikis, the normal checkpoint program of programmed cell death – a key prerequisite step for cancer metastasis. Systematic causal investigation of shape and signaling is notoriously difficult experimentally whereby perturbation may destroy the intricate feedback between the two, and/or results may be challenging to interpret. In this talk, I will present the development of alternative non-perturbation computational approaches, inspired by statistical inference techniques used in financial mathematics, neuroscience and ecology to quantify causal relationships between local 3D cell surface morphology and membrane-associated signals in a spatiotemporal manner. Specifically, I will discuss the development of three core computational frameworks that I have developed: u-InfoTrace for computing spatiotemporal information flows; u-Segment3D for precise 3D cell segmentation; and u-Unwrap3D for performing generalized computing on rugged 3D surfaces. I will show applications of using these tools to build workflows for analyzing high-resolution microscopy videos of cell blebbing, cancer-immune cell interactions, and organoids.