Stages 2023
Stage 1 : Machine learning for immune dynamics
A typical immune response is multiplexed and highly dynamical : after exposure to a pathogen, cytokines are secreted and multiple cells are activated. Recently, we identified a “universal antigen encoding” phenomenon, revealing how the immune system encodes information in the dynamics of the cytokines (see movie on the right). But one does not know how typical immune cells interpret that information. The goal of this stage is to use machine learning to analyze data provided by our collaborators at NIH, to understand how cells decode the universal antigen encoding.
Stage 2 : Modelling Clocks and Unclocks in Embryonic development
During embryonic development, waves of genetic expressions sweep across the embryo to define and localize future vertebrae. Those waves of oscillations define the so-called “segmentation clock” . But is the segmentation clock a “true” clock ? Winfree categorized biological oscillators into clocks and unclocks, the goal of the stage will be to study (numerically and analytically) models of biological oscillators to study how their properties can change from “clock” to “unclock”, to eventually relate them to the properties of the segmentation oscillator.