UC San Diego

Bottom-up, genome-scale metabolic reconstructions have been used to model the growth phenotypes of E. coli in various conditions. Experimental discrepancies with model predictions indicate 1) modeling failed to predict an achievable state of E. coli due to errors in model parameters or network structure; OR 2) adaptive evolution is required to achieve the computationally predicted state. The modeling predictions of the iAF1260 E. coli metabolic reconstruction were tested experimentally for over 4,000 scenarios, uncovering both improvements to be made to network structure and hypotheses for the discovery of novel isozymes and pathways in E. coli. Furthermore, the genetic and molecular basis for adaptation to a computationally predicted growth state was determined through whole-genome resequencing of eleven strains adapted to growth in lactate M9 minimal medium and through detailed studies of highly-adaptive RNA polymerase mutants at the kinetic, physiological, and systems biology levels