Figure 2

Engineering meta-phenotypes for economic data set. Each column (listed by Cluster Id and grouped by organism) is the centroid associated with a corresponding k-means phenotype cluster (meta-phenotype) for the simulation data subset with economic metrics. Cluster Ids increase sequentially, from left to right: 1 to 10 for E. coli, 11 to 30 for S. cerevisiae, and 41 to 60 for S. oneidensis. To compare across phenotype clusters and organisms, the metric values have been transformed into row-wise z-scores. All the rows and organism-specific columns were then hierarchically clustered. The “cluster sizes” are the number of individual phenotype simulations associated to each phenotype cluster. The “cluster heterogeneity” is the within-cluster sum of squared errors (SSE) for each phenotype cluster normalized by the maximum for each organism.