平特五不中

With the advance of machine learning, there has been a renewed interest in the inference and the study of the geometry underlying biological dynamics. I will present two recent applications of such ideas, where our group could combine theory and experiment to build predictive models of complex biological processes. In the context of embryonic development, experimental data on the entrainment/coupling of the segmentation oscillator, and new geometric models of somitogenesis, suggest the existence of an asymmetric core oscillator (work in collaboration with Alexander Aulehla, EMBL Heidelberg). In the context of immune response, we derived explicitly a simple geometric picture to describe complex cytokine dynamics, revealing a universal "antigen encoding", with applications for CAR-T cell based immunotherapy (work in collaboration with Gr茅goire Altant-Bonnet, NIH Bethesda). Our work reveals how systems level descriptions can be built from data, leading to surprisingly accurate descriptions and associated predictions.

This seminar will be offered online via Zoom. Details in attached poster.