Statistical Physics of Information Processing by Cells
This event is part of the Departmental Seminars.
Dissertation Committee: Pankaj Mehta, Kirill Korolev, Claudio Chamon, Andrew Liam Fitzpatrick, Allyson Sgro
Eukaryotic cells transmit information using complex biochemical signaling networks based on protein-protein interactions (PPIs). In this talk, I'll discuss a theoretical and computational framework I developed which relates network-level information processing properties to the biophysical properties of PPIs. To do so, I first generalize statistical physics-inspired models for protein binding to account for post-translational modifications (PTMs) of proteins (e.g., phosphorylation) and PTM-dependent binding, both of which are essential characteristics of signaling. By combining these models with information theoretic methods, I’ll show that the strength of PPIs acts as a regulator of information transmission: weak PPIs give rise to “noise” that limits information transmission and that such noise mediates a fundamental trade-off between speed of dynamical response and information. I’ll also discuss a surprising finding that cross-talks between pathways in complex signaling networks do not significantly alter information capacity, an observation that may partially help explain the promiscuity and ubiquity of weak PPIs in naturally-occurring signaling networks. Finally, I’ll conclude by showing how these ideas can be used to design PPIs in synthetic biochemical networks to maximize information transmission, a procedure we dub "InfoMax" design.