2025/04/12 by Lahr, Pedro Henrique
#Bioinformatics #Biotechnology #Codon Optimization #Computational Biology #Computational Modeling #FOS: Computer and information sciences #Gene Expression #Gene Synthesis #Genetic Engineering #Genetics and Genomics #Genomic Design #Life Sciences #Mathematical Biology #Molecular Biology #Pre-symptomatic Testing #Protein Expression Engineering #Proteomics #Synonymous Codon Usage #Synthetic Biology #Systems Biology #Systems Modeling
paper · doi:10.17605/osf.io/d8wqh
Stochasticity, or noise, is an intrinsic feature of gene expression, leading to cell-to-cell variability in protein levels even within isogenic populations. This variability plays critical roles in biological processes ranging from microbial bet-hedging to cellular differentiation, yet it also poses challenges for the predictable engineering of synthetic biological systems. While controlling the mean level of protein expression is well-established, strategies to independently tune expression noise remain less developed. This project investigates the hypothesis that the sequence of synonymous codons within a gene's coding region can be rationally designed to modulate protein expression noise, potentially decoupling it from the mean expression level. The underlying mechanism is proposed to involve the alteration of ribosome traffic dynamics – specifically, how patterns of "fast" and "slow" translating codons influence ribosome pausing, queuing, and translational bursting. Purpose and Approach: The primary purpose of this study is to establish a computational framework for predicting how different synonymous codon usage strategies impact protein expression noise and to provide a theoretical basis for using codon design as a noise-tuning tool in synthetic biology. Given the computational demands of explicitly simulating detailed ribosome traffic (e.g., using TASEP models), this initial study employs a computationally efficient, simplified two-stage stochastic model of transcription and translation. Within this framework, the effect of different codon usage strategies (e.g., fully optimized using high-tAI codons, deoptimized using low-tAI codons, or specific patterns like slow initial ramps or central bottlenecks) is represented by distinct effective translation rates per mRNA (ktlvariant). We performed stochastic simulations using the Gillespie algorithm (optimized with Numba, code provided in Appendix) to analyze the resulting steady-state protein distributions for each representative ktlvariant. Key Findings & Expected Outcomes: Our simulations using the simplified model confirm that the effective translation rate significantly impacts both the mean protein level () and standard noise metrics. We observed the expected inverse relationship between mean expression and relative noise (Coefficient of Variation, CV), with higher translation rates leading to lower CV (e.g., CV ≈ 0.07 for a "Fast" strategy proxy). Importantly, the simulations also revealed significant translational burstiness (Fano Factor ≈ 39 for the "Fast" proxy), highlighting the inherent stochasticity captured even by this simplified model due to mRNA lifetime dynamics. While this computationally tractable model validates the fundamental link between average translation parameters (influenced by codon choice) and expression statistics, it does not explicitly model ribosome traffic jams and thus cannot directly test the hypothesis of Mean-Noise decoupling via specific codon arrangements. We discuss the TASEP-based mechanistic hypothesis that predicts such decoupling is possible and propose that engineering codon patterns (like bottlenecks) could allow fine-tuning of noise independently of the mean in vivo. The main outcome of this work is a computational proof-of-principle, a lightweight simulation code, and, crucially, a detailed experimental plan designed to rigorously test the core hypothesis using fluorescent reporters and single-cell flow cytometry in E. coli. Successful experimental validation would establish synonymous codon design as a novel, powerful tool for the rational engineering of gene expression dynamics with desired stochastic profiles, impacting the design of robust synthetic circuits and enabling precise control over cellular heterogeneity. This preprint manuscript details the background, the simplified computational model and its results, the underlying mechanistic hypothesis involving ribosome traffic, and the comprehensive experimental validation strategy.