Automating experimental procedures has resulted in an unprecedented increase in the volume of generated data, which, in turn, has caused an accumulation of unprocessed data. As a result, the need to develop tools to analyze data systematically has been rising. Also, the growing computational power of computers has enabled the development of more advanced tools to understand and engineer intricate systems. Biological networks are one such intricate system consisting various interplaying components. Systems biology concerns developing databases, models, and methods to organize, simulate, and analyze the bulk of data about biological networks, including metabolism. Tremendous progress in gene sequencing and annotation on one side, and the advent of automatic tools to translate gene annotations into a network of reactions catalyzed by the gene products on the other side, have enabled the reconstruction of thousands of metabolic models at the genome scale. Assuming cellular metabolism converges to a steady state to pursue an objective, metabolism can be simulated at steady-state without the need to account for variations over time. Due to these advantages, metabolism has been extensively studied and simulated. This has rendered metabolism an appropriate scaffold for integrating other biological processes and reconstructing expanded models. In this thesis, we reconstruct models and devise methods to study cellular processes, where we begin with metabolism and expand the scope. First, we review the objective functions used in the study of metabolism, which provides insight into the design principles of metabolic networks. We follow by presenting methods to simulate and engineer single-cell metabolism. In particular, we present two methods. SubNetX merges graph-search algorithms and constraint-based optimization to suggest optimal pathways to synthesize a valuable compound; TOptKnock exploits bilevel optimization to search for thermodynamically feasible knockouts to couple biomass and product yields. Then, we widen the scope and hence increase the complexity of our studies in two directions. First, we integrate more biological processes, specifically protein expression, reconstruct models of metabolism and expression for Escherichia coli and Saccharomyces cerevisiae, and apply these models to simulate recombinant protein expression. Second, we account for metabolic interactions between different cells with a focus on microbial communities. To this end, we specify the metabolic potential of individual species to uptake nutrients and secrete byproducts. Then, we use such information to reconstruct interaction networks while optimizing an objective function. We can reconstruct interactions based on a community driving force, partial information about the interactions, or an engineering goal using various objective functions. The methods and models presented in this thesis demonstrate how systems biology helps understand biological processes, optimize experiments, and design synthetic organisms.