Genome-Scale Models of Metabolic and Regulatory Networks - Systems Biology: Networks, Models, and Applications - page 253

Figure 8.5 Iterative model-building process for regulated metabolic models.
An intial model (dark lines including regulation, lighter lines corresponding to
unregulated model) is used to define an allowable solution space. Experimental
data is compared to model predictions and discrepancies are identified. The
model is then expanded to improve agreement with data and tested against an
independent data set to avoid overfitting. The process is then repeated by
designing a new set of experiments to further expand the model.
powerful way to bring together multiple types of high-throughput
data (e.g., gene expression and phenotyping) and to interpret these
data sets. Discrepancies between model predictions and experi-
mental data can be used to iteratively improve the regulatory network
reconstruction.
The approaches described above aim to construct a model of the
regulatory network based on, for example, ChIP-chip and gene expres-
sion data and then combine this model with the genome-scale
metabolic model. An alternative to this approach is to use gene expres-
sion data directly as an additional constraint in the analysis of metabolic
networks on a condition-by-condition basis [110]. Constraining fluxes
through downregulated reactions in a genome-scale model of yeast
was shown to result in improved prediction of exchange and intracel-
lular fluxes [110]. Since the relationship between metabolic flux and