Systems biology is a rapidly growing discipline. It is widely believed to have a broad transformative potential on both basic and applied studies in the life sciences. In particular, biochemical network reconstructions are playing a key role as they provide a framework for investigation of the mechanisms underlying the genotype- phenotype relationship. In this thesis, the procedure to reconstruct metabolic networks is illustrated and extended to other cellular processes. In particular, the constraint -based reconstruction and analysis approach was applied to reconstruct the transcriptional and translational (tr/tr) machinery of Escherichia coli. This reconstruction, denoted 'Expression-matrix'/ (E-matrix), represents stoichiometrically all known proteins and RNA species involved in the macromolecular synthesis machinery. It accounts for all biochemical transformations to produce active, functional proteins, tRNAs, and rRNAs known to be involved in macromolecular synthesis in E. coli. An initial study investigated basic properties of the E- matrix, including its capability to produce ribosomes, which was found to be in good agreement with experimental data from literature. Furthermore, quantitative gene expression data could be integrated with, and analyzed in the context of, the resulting constraint-based model. Adding mathematically derived constraints to couple certain reactions in the model allowed the quantitative representation of the size of steady state protein and RNA pools. Furthermore, the E-matrix was integrated with the genome-scale E. coli metabolic model and extended the transcriptional and translational reactions to encompass genes encoding all the respective metabolic enzymes. The resulting Metabolite-Expression-matrix (MExv matrix), has exceeds the predictive capacity of the metabolic model and it can, for example, be used to predict the biomass yield since it represents the production of almost 2,000 proteins. E. coli 's ME-matrix is the first of its kind and represents a milestone in systems biology as demonstrates how to quantitatively integrate 'omics'- datasets into a network context, and thus, to study the mechanistic principles underlying the genotype-phenotype relationship. Possible applications are just beginning to become apparent and may include protein engineering, interpretation of adaptive evolution, and minimal genome design. An integration of the ME-matrix with remaining cellular processes, such as regulation, signaling, and replication, will be a next step to complete the first whole-cell model.