Stochastic trade-offs and the emergence of diversification in E. coli evolution experiments

TL;DR

This paper investigates the stochastic factors influencing evolutionary diversification in E. coli, introducing stochastic trade-offs to better explain variability in experimental outcomes and bridging the gap between theory and empirical observations.

Contribution

It introduces the concept of stochastic trade-offs in trait space to explain variability in evolutionary branching, extending adaptive dynamics models.

Findings

Stochastic trade-offs account for observed variability in diversification.

Predictions of branching likelihood align with experimental data.

Variability is constrained to a flexible tradeoff curve.

Abstract

Laboratory experiments with bacterial colonies, under well-controlled conditions often lead to evolutionary diversification, where at least two ecotypes emerge from an initially monomorphic population. Empirical evidence suggests that such ''evolutionary branching'' occurs stochastically, even under fixed and stable conditions. This stochastic nature is characterized by: (i) occurrence in a significant fraction, but not all, of experimental settings, (ii) emergence at widely varying times, and (iii) variable relative abundances of the resulting subpopulations across experiments. Theoretical approaches to understanding evolutionary branching under these conditions have been previously developed within the (deterministic) framework of ''adaptive dynamics''. Here, we advance the understanding of the stochastic nature of evolutionary outcomes by introducing the concept of ''stochastic trade-offs'' as opposed to ''hard'' ones. The key idea is that the stochasticity of mutations occurs in a high-dimensional trait space and this translates into variability that is constrained to a flexible tradeoff curve. By incorporating this additional source of stochasticity, we are able to account for the observed empirical variability and make predictions regarding the likelihood of evolutionary branching under different conditions. This approach effectively bridges the gap between theoretical predictions and experimental observations, providing insights into when and how evolutionary branching is more likely to occur in laboratory experiments.