(582r) Harnessing Flux Balance Analysis for Simulating Biological Evolution

Current evolution simulators rely on artificial fitness landscapes that are meant to represent a collection of genotypes that may be generated through random mutation during population growth. Despite the many successes of these tools and other theoretical approaches that make similar approximations, a representative genome that allows for direct linkage between metabolism in silico and in vivo as the basis for simulated evolution would provide experimenters with a powerful new method to predict mutations of interest prior to experimental evolution. To address this issue, we have developed a genetic algorithm that represents individual cells using metabolic flux models, providing a direct linkage between genotype and metabolic behavior of all mutants during in silico evolution. After demonstrating the functionality of the simulator with an example model, gene expression data for Escherichia coli under n-butanol and isobutanol stress is utilized to identify possible compensatory, non-trivial metabolic changes that may occur during adaptive evolution to increase biofuel tolerance. Previously collected experimental data were found to agree with the simulator predictions, demonstrating the effectiveness of this approach for strain engineering purposes.