Skip to main content
Fig. 1 | BMC Systems Biology

Fig. 1

From: Integrating transcriptional activity in genome-scale models of metabolism

Fig. 1

The CoRegFlux workflow, in which a training gene expression data set is used to infer a coregulatory network (in this case from the M3D database [19]). Using this inferred network, we calculated the influence scores of the regulators (i.e. TFs, kinases). We then used these scores to train a linear model for predicting gene expression from influence scores. This model can use context-specific influence scores to predict the activity of metabolic genes in a robust manner. Using these predictions, the model employs a continuous-value version of the Gene-Protein-Reaction rules and a function to translate gene activity into flux bounds. The bounds obtained are then input into a linear program to obtain fluxes congruent with the gene-regulatory state of the cell in a given context

Back to article page