New computational framework predicts whether engineered microbial traits survive evolution

Engineered microbial strains used in industry often face an uncomfortable reality: as their populations expand, Darwinian selection pushes them to evolve, and the carefully developed traits that make them valuable can erode. A new study in PLOS Computational Biology introduces TRAEV, short for Trait Robustness Against EVolution, a computational framework designed to predict in silico whether desired metabolic traits will persist as microbes adapt to their application environments.

The framework was developed by Chunhao Gu, Ville Mustonen and Paula Jouhten. It builds on constraint-based metabolic models, which allow researchers to predict how environment-dependent traits relate to fitness, and it uses Monte Carlo-based perturbation analysis to capture the stochastic nature of adaptive evolution.

TRAEV works in two stages. First, it predicts the immediate metabolic phenotype of a strain when placed in a new environment. It then simulates the subsequent evolutionary trajectories of both fitness and the desired trait, such as the production of a flavor compound or a platform chemical.

From these simulated trajectories, the framework derives two quantitative measures. The Robustness Score, or RS, represents trait robustness and the overall trait level during evolutionary adaptation. The Trade-off Score, or TS, captures trade-offs between the trait and fitness, revealing cases where maintaining a desired trait comes at a cost to the organism’s growth.

To validate the approach, the researchers predicted the relative evolutionary robustness of heterologous pigmentation in two engineered strains of the yeast Saccharomyces cerevisiae, as well as two desired aroma generation traits in multiple wine strains of the same species, assuming the traits were developed through a laboratory selection process. The results showed how TRAEV predictions could guide strain development decisions.

The problem the framework addresses is economically significant. Microbial strains undergo notable biomass expansion in industrial applications, subjecting populations to Darwinian selection. The consequent adaptive evolution may threaten desired metabolic traits that are often tediously developed, and until now it remained unresolved how to predict the evolutionary robustness of such traits.

Because TRAEV is built on constraint-based modeling, it can be integrated into existing computational strain design workflows across different microbial strains, desired metabolic traits and application environments. This makes it a flexible tool rather than a solution tailored to a single organism or product class.

The authors note that model-predicted evolutionary robustness of desired metabolic traits can ultimately guide both strain and process development, helping to enhance the economic attractiveness of industrial applications that rely on microbial cells.

Subject of Research: Computational prediction of the evolutionary robustness of engineered metabolic traits in microbial cells

Article Title: In silico prediction of metabolic trait robustness in microbial cells

Article References: Gu, C., Mustonen, V., & Jouhten, P. (2026). In silico prediction of metabolic trait robustness in microbial cells. PLOS Computational Biology, 22(10), e1014864. https://doi.org/10.1371/journal.pcbi.1014864

Image Credits: AI Generated

DOI: 10.1371/journal.pcbi.1014864

Keywords: TRAEV, metabolic engineering, adaptive evolution, constraint-based models, Saccharomyces cerevisiae, trait robustness, fitness trade-offs, Monte Carlo simulation, industrial biotechnology, strain design, PLOS Computational Biology, silico