Genome-Scale Models of Metabolic and Regulatory Networks - Systems Biology: Networks, Models, and Applications - page 254

gene expression patterns appears to be quite complicated [111], further
advances in integrated metabolic/regulatory network modeling
methods will be needed [112-115]. Constraint-based modeling provides
a good framework for developing such methods through successive
application of more complex regulatory constraints.
CONCLUSIONS AND FUTURE DIRECTIONS
Reconstruction of genome-scale biochemical networks is necessary
for understanding integrated network properties and for building
quantitative predictive models of these networks. Genome sequences
as well as extensive biochemical literature have allowed reconstruc-
tion of genome-scale metabolic reaction networks for many microbial
organisms. Methods have been developed to allow constraint-based
analysis of the resulting stoichiometric matrices in order to predict
experimentally measurable physiological parameters. These predic-
tions have been shown to agree quantitatively with experimental
measurements for evolved strains and qualitatively with growth
phenotyping data for nonevolved knockout strains. In addition, con-
straint-based analysis of genome-scale metabolic networks has allowed
studying in silico global properties such as redundancy and robustness
of these networks. Recent developments in in silico methods have also
allowed utilizing genome-scale models in metabolic engineering [72,74,76]
and other practical applications in, for example, bioremediation are
already emerging [116].
With the increased availability of high-throughput experimental data,
the reconstruction of transcriptional regulatory networks is becoming
feasible at the genome scale, for any organisms that has been suffi-
ciently well characterized and for which suitable data is available.
Specific properties of regulatory networks, such as lack of evolutionary
conservation of binding sites, make both reconstruction and modeling
of these networks more challenging than the corresponding tasks for
metabolic networks. New developments in experimental techniques,
data-based reconstruction approaches, and in silico modeling methods
are still needed to improve our ability to build quantitative models
of regulatory networks. Nevertheless, the first large-scale regulatory
network models that have been reconstructed show that we are already
in the position to build integrated models that incorporate both reg-
ulatory and metabolic networks [57,102]. These models allow
simultaneous analysis of multiple data types as well as the studying
the effects of the interplay between metabolic and regulatory networks
on overall cellular function.
Genome-scale models of cellular processes, other than metabolism
or transcriptional regulation, are also beginning to be developed.
In particular, sufficient amounts of information on signaling networks