< Back to Article

Comparative Analysis of Yeast Metabolic Network Models Highlights Progress, Opportunities for Metabolic Reconstruction

Fig 5

Model prediction of single-gene essentiality is sensitive to biomass definition.

Since objective function is a tunable model parameter, we calculated Matthews’ Correlation Coefficients for the sum of all true positive, true negative, false positive and false negative predictions across all conditions using two different objective functions for each model: the biomass definition provided by the model authors, and the biomass function used for the iFF708 model. We found that with the exception of the Yeast 4 model, all model predictions were improved by tuned objective function, independent of refinements to the biochemical network reconstruction. Models are arranged in chronological order across the horizontal axis.

Fig 5