Non equilibrium dynamics in Escherichia coli's gene regulatory network

Gene regulation is a key process in living organisms. It defines cells identity and behavior, and allows the cells to adapt to the external environment.
From a theoretical point of view, a central question is how to mathematically characterize the many players and their complex interactions to understand the gene expression output as function of the regulatory inputs.
A common approach is thermodynamic modelling, where the transcription is assumed to be in equilibrium with the concentration of transcription factors, and any fluctuation is averaged away. However, the advent of new experimental techniques providing precise measurements of gene expression at the single-cell level is challenging the general validity of the equilibrium assumption.
In this thesis, we focus on the induction of the LexA regulon in the model organism Escherichia coli, which is involved in the repair of DNA damages. Tracking the single-cell expression dynamics of different genes under the exclusive control of the inhibitor LexA, we show that the induction is characterized by short bursts of production, which are incompatible with a thermodynamic model where gene transcription is in equilibrium with the concentration of LexA. On the other hand, we show that the network responds to transient fluctuations in the concentration of the regulator.
Finally, we deal with the question of how to properly analyze flow cytometry data for bacterial populations. Flow cytometry is an attractive technology to quantify single-cell gene distribution in high-throughput. However, so far no systematic investigation has been carried out to estimate the accuracy of these measurements for small bacterial cells. Here, by comparing the fluorescence distribution of the same E. coli strain both in flow cytometry and in a microscopy setup, we show that the fluorescent signal contains a significant amount of electronic noise and background fluorescence. We then propose a robust method to correct for these spurious components, and we show that only after correcting for electronic noise and autofluorescence, measurements from the flow cytometry agree with the ones from the microscopy setup.