Figure 3

Pareto optima and trade-offs for multiple engineering goals. (A-C). Two-dimensional (2D) and three-dimensional (3D) candidate and Pareto optimal designs and frontiers for engineering goals related to succinate production. Sub-optimal candidate designs for all organisms and experimentally validated E. coli designs are shown in (A) and (B), but are not shown in (C) for clarity. (D-F). 2D and 3D candidate and Pareto optimal designs that had matching experimental microarray data and frontiers for engineering goals related to acetate production. Error bars are too small to be visualized. For comparison, candidate designs that use the same media conditions as the microarray designs are shown in (D), however only the microarray designs are used to compute Pareto optima and frontiers. Sub-optimal microarray designs are shown in (D) and (E), but are not shown in (F) for clarity. The 3D Pareto frontiers shown in (C) and (F) are colored in proportion to the total economic cost and acetate production rates, respectively. Trade-offs can be determined by evaluating the rate of change of the Pareto frontier over the desired range of Pareto optimal designs. “Validated” designs are designs that were experimentally validated, “microarray” designs are designs that had matching experimental microarray data, and “candidate” designs are simulated designs that are neither “validated” nor “microarray” designs. Abbreviations: ec = Escherichia coli, sc = Saccharomyces cerevisiae, so = Shewanella oneidensis. The legend on top of the panels summarizes the symbols used.