Stochastic single-gene auto-regulation

TL;DR

This paper develops a comprehensive stochastic model of single-gene auto-regulation, analyzing how intrinsic randomness affects protein and mRNA distributions under different timescale regimes, with implications for understanding gene expression variability.

Contribution

It introduces a detailed stochastic model that includes all sources of intrinsic noise and derives analytic solutions for equilibrium distributions in different timescale regimes.

Findings

Unimodal and bimodal protein distributions can both occur.

The model's analytic solutions agree well with simulations.

Bimodality is linked to specific auto-regulation mechanisms.

Abstract

A detailed stochastic model of single-gene auto-regulation is established and its solutions are explored when mRNA dynamics is fast compared with protein dynamics and in the opposite regime. The model includes all the sources of randomness that are intrinsic to the auto-regulation process and it considers both transcriptional and post transcriptional regulation. The timescale separation allows the derivation of analytic expressions for the equilibrium distributions of protein and mRNA. These distributions are generally well described in the continuous approximation, which is used to discuss the qualitative features of the protein equilibrium distributions as a function of the biological parameters in the fast mRNA regime. The performance of the timescale approximation is assessed by comparison with simulations of the full stochastic system, and a good quantitative agreement is found for a wide range of parameter values. We show that either unimodal or bimodal equilibrium protein distributions can arise, and we discuss the auto-regulation mechanisms associated with bimodality.