Gene network robustness as a multivariate character

Abstract

Robustness to genetic or environmental disturbances is often considered as a key property of living systems. Yet, in spite of being discussed since the 1950s, how robustness emerges from the complexity of genetic architectures and how it evolves still remains unclear. In particular, whether or not robustness to various sources of perturbations is independent conditions the range of adaptive scenarios that can be considered. For instance, selection for robustness to heritable mutations is likely to be modest and indirect, and its evolution might result from indirect selection on a pleiotropically-related character (e.g., homeostasis) rather than adaptation. Here, I propose to treat various robustness measurements as quantitative characters, and study theoretically, by individual-based simulations, their propensity to evolve independently. Based on a simple evolutionary model of a gene regulatory network, I showed that different ways to measure the robustness of gene expression to genetic or non-genetic disturbances were substantially correlated. Yet, robustness was evolvable in several dimensions, and robustness components could evolve differentially under direct selection pressure. Therefore, the fact that the sensitivity of gene expression to e.g. mutations and environmental factors rely on the same gene networks does not preclude that robustness components may have distinct evolutionary histories.