Abstract

Theoretical biology has often relied on physics-inspired metaphors (such as the Waddington landscape). However, it is not clear if and how such ideas can be used to describe the dynamics of complex gene networks. In this talk, I will show how signalling dynamics can be represented and modelled in a low dimensional « latent space », offering a complementary approach to classical gene network modeling. I will illustrate this strategy with two examples. We studied dynamics of vertebrae formation and showed how geometry constraints specific bifurcation patterns, connecting many aspects of the data. In an immune context, we developed a robotic platform combined with machine learning to derive a « ballistic » model of cytokine dynamics, explaining common patterns of immune activations.

About the speaker

Dr. Paul François is an Associate Professor in the McGill University Physics Department. His group researches theoretical biological physics and work on such topics as the evolution of gene regulatory networks, cellular detection in the immune system and vertebrate development.

Website: https://www.mcgill.ca/francois-group/