Microorganisms are essential for life on Earth, performing key roles in numerous biological processes. Their influence extends across a wide spectrum, from human health and ecological balance to advancements in biotechnology and industrial applications. These powerful entities do not live in isolation but form complex communities, known as microbiomes, integral to the functioning of natural and engineered environments. Despite their importance, our understanding of how these tiny organisms form and evolve in nature is limited. Metabolism, the network of life-sustaining enzyme-catalyzed biochemical reactions within organisms, is fundamental to the evolution and functioning of microbiomes. From the microscale of cellular enzymes to the broader context of microbial communities, metabolic processes are an outcome of evolution and are thus shaped by certain design and optimality principles. Over the last decades, mathematical and computational models have been pivotal in deciphering these principles. More recently, with the substantial growth of omics data, coupled with advancements in high-throughput technologies, there is a growing necessity to develop computationally efficient frameworks that can shed light on the underlying design principles of metabolism across various scales. In this thesis, we present computational methods to understand the design and optimality principles in biological processes. We address the mathematical and biological challenges involved in their formulation and the uncertainties they introduce. Firstly, we investigated cellular enzymes through an evolutionary approach and introduced the OpEn framework to estimate optimal modes of operations of cellular enzymes, applicable to any complex enzyme mechanism. We then expanded our focus to encompass metabolic networks, devising a unifying framework for studying dynamic metabolic behavior in microorganisms. This approach addressed the uncertainty associated with the multiplicity of solutions and showed how different optimality principles and mathematical formulations lead to different dynamic trajectories. Moving to microbial communities, we introduced the ReMIND workflow for reconstructing metabolic interaction networks. Utilizing ReMIND, we investigated a range of objective functions, explored the trade-offs between metabolite exchange and biomass yields, identified metabolite hubs with high connectivity, and evaluated the effects of adding or removing species on metabolic interactions. Lastly, we investigated the spatial organization of a synthetic microbial community during colony expansion and studied the effects of environmental and biochemical perturbations on the emerging spatial patterns. Overall, the research in this thesis encompasses a broad spectrum, from cellular enzymes to microbial communities, and extends from steady-state conditions to temporal and spatial dynamics. We introduce mathematical models and methodologies that investigate and clarify the intricate design and optimality principles governing biological processes across different scales. This comprehensive exploration enhances our understanding of microbial emergence and evolution in diverse contexts.