The ability to predict mutation spread in engineered cell populations has important implications for synthetic biology research and industrial bioproduction. Despite this, existing tools to predict spread are limited to simplistic descriptions of mutation and selection. Accurately accounting for mutation dynamics would instead require linking the design of a synthetic gene construct to (i) its mutation probability and (ii) its impact on population-wide selection. These elements have been studies in isolation, but have not been combined into an analysis of mutation spread.

Motivated by this gap, we combine a model of state transitions with an existing whole-cell model to show how a synthetic construct's gene expression affects the consumption of shared resources, thereby impacting cell growth rate and selection rate in a growing population. We expand this framework to consider how mutation heterogeneity impacts mutation spread, finding that both the 'severity' of a mutation and its 'location' on a synthetic construct are relevant when capturing experimentally-observed dynamics. We demonstrate the potential of the framework by applying it to a synthetic repressilator and explore the role of mutation heterogeneity on its robustness. We show that completely disabling genes, as opposed to partially reducing their function, is selectively disadvantageous because the total burden from synthetic gene expression increases. This suggests that constricting possible mutation pathways increases a repressilator's robustness, a feature that has been previously observed experimentally, but not explained. Finally, we develop a tool to predict a synthetic construct's mutation probability from its sequence using literature data on mutation rates, taking inspiration from the existing tool 'EFM calculator'. We improve on this by (i) separating the effects of RecA-dependent and RecA-independent homologous recombination, (ii) using more complete links between between sequence features and mutation mechanisms, and (iii) considering the effect of methylation on the base pair substitution rate.

In total, by combining the effects of selection, mutation heterogeneity and mutation probability, we provide a novel modelling framework that is well-suited to the varied construct designs of present-day synthetic biology.