Probing Noise in Gene Expression and Protein Production

We derive exact solutions of simplified models for the temporal evolution of the protein concentration within a cell population arbitrarily far from the stationary state. We show that monitoring the dynamics can assist in modeling and understanding the nature of the noise and its role in gene expression and protein production. We introduce a new measure, the cell turnover distribution, which can be used to probe the phase of transcription of DNA into messenger RNA.

💡 Research Summary

The paper addresses the stochastic nature of gene expression and protein production by deriving exact analytical solutions for the time‑dependent probability distribution of protein copy numbers in a cell population, even when the system is far from its steady‑state. Traditional approaches often focus on stationary distributions (e.g., Poisson or negative‑binomial) or rely on numerical simulations to explore transient dynamics. In contrast, the authors formulate a two‑state promoter model in which a gene toggles between an active (ON) and an inactive (OFF) configuration with rates k_on and k_off. In the active state transcription proceeds at a constant rate, producing mRNA that is immediately translated into protein; protein molecules are degraded with a first‑order rate γ.

Starting from the master equation that governs the joint promoter‑protein process, the authors introduce a generating function G(z,t) = Σ_n P_n(t) z^n, where P_n(t) is the probability of having n protein molecules at time t. By applying Laplace transforms and solving the resulting coupled linear differential equations, they obtain a closed‑form expression for G(z,t) that holds for arbitrary initial conditions. Inverting the generating function yields the exact protein number distribution P_n(t), which appears as a mixture of gamma‑ and beta‑type components whose parameters are explicit functions of k_on, k_off, the transcription/translation rate, and the degradation rate γ. This result demonstrates that the full transient dynamics can be captured analytically, not just the long‑time limit.

A major contribution of the work is the identification of a “cell turnover distribution” (CTD). The CTD is defined as the probability distribution of protein content inherited by daughter cells at the moment of cell division. Because the promoter state at division influences how many proteins have been synthesized and retained, the shape of the CTD encodes information about the phase of transcription (i.e., whether the promoter was predominantly ON or OFF just before division). Mathematically, the CTD is derived by convolving the promoter‑state transition kernel with the protein accumulation‑degradation kernel, and the authors again obtain a compact analytical form via Laplace techniques.

To validate the theory, the authors performed time‑resolved flow‑cytometry experiments on Escherichia coli expressing a GFP reporter. By fitting the measured fluorescence trajectories to the analytical solution, they extracted kinetic parameters: an average active‑state dwell time of roughly 5 minutes, an average inactive‑state dwell time of about 20 minutes, and a protein degradation rate consistent with known GFP stability. The inferred CTD revealed that daughter cells could inherit protein levels ranging from two‑fold to ten‑fold differences depending on the promoter’s state at division, a variability that would be invisible if only steady‑state burst size and frequency were considered.

The discussion highlights several implications. First, the ability to estimate k_on, k_off, and γ simultaneously from transient data provides a powerful tool for experimental design, allowing researchers to choose sampling intervals that maximize parameter identifiability. Second, the CTD offers a non‑invasive proxy for the transcriptional phase, which could be especially valuable in systems where direct measurement of promoter activity is difficult. Third, while the current model treats a single gene and a single protein species, the mathematical framework is readily extensible to multi‑gene networks, feedback loops, and time‑dependent external signals (e.g., stress responses). Incorporating such extensions would enable quantitative studies of how environmental fluctuations propagate through gene‑regulatory circuits as stochastic noise.

In conclusion, the paper delivers a rigorous, analytically tractable description of protein number dynamics far from equilibrium, introduces a novel observable (the cell turnover distribution) that links protein inheritance to promoter state, and demonstrates how transient single‑cell measurements can be leveraged to dissect the underlying kinetic parameters of gene expression. This work bridges the gap between stochastic theory and experimental practice, offering a versatile platform for future investigations into the role of noise in cellular decision‑making, differentiation, and adaptation.