Accèder directement au contenu
7 avril

Lev Tsimring Information capacity of dynamical signaling networks

11h30 à 12h30

Le séminaire de Lev Tsimring (Biocircuits Institute, UCSD) aura lieu dsalle GH222, 2e étage du département de physique de l’ENS (entrée : 18 rue Lhomond)

The fidelity of intracellular signal transduction is limited by the stochasticity of underlying biochemical reactions (intrinsic noise) and variability of the cellular state (extrinsic noise). Previous studies that have focused on estimating information capacity as mutual information between a scalar input signal and a scalar output response, found a very limited capacity of many biological networks (usually less than 1 bit). In this work we demonstrate theoretically and experimentally that dynamical signaling networks allow for a significantly higher information capacity than static networks. Our information-theoretical analysis shows that signaling dynamics helps greatly reduce or even eliminate losses due to extrinsic noise in addition to averaging out the intrinsic noise. We confirm our theoretical results by measuring channel capacity of three important signaling pathways, Erk, Calcium, and NFkB. To that end, we developed a new binless algorithm for estimating mutual information between a scalar input and a vector output directly from data. Our work demonstrates that cells can use dynamics of signaling networks to overcome the degrading effects of extrinsic noise and increase accuracy and variety of cellular responses to environmental cues.