< Back to Article

Inferring Metabolic States in Uncharacterized Environments Using Gene-Expression Measurements

Figure 5

Comparing the quality of gene essentiality predictions.

The sensitivities and positive predictive values (PPV) of the prediction of essential genes are shown for YPD and YPEtOH. The methods compared are EXploration of Alternative Metabolic Optima (EXAMO), Highly Expressed Genes as proxies for essential genes (HEG), Shlomi et al's integrative Metabolic Analysis Tool (iMAT, Eqs. 16) [27], and Jerby et al. 's Model Building Algorithm (MBA) [26]. The blue horizontal line marks the sensitivities and PPVs achieved with the yeast model when the free uptake and secretion of all exchange metabolites is allowed. These genes predicted to be essential for growth using the original yeast model, unconstrained by gene expression measurements, are predicted to be essential regardless of which nutrients are available to the cells.

Figure 5