A majority of protein function is mediated through direct binary or complex interactions with other proteins. Therefore, systematic efforts to characterize these protein interaction networks and to structurally resolve their interaction interfaces have provided powerful tools to comprehensively study protein function at a molecular level. For instance, disease mutations are enriched along protein interaction interfaces and network level impacts of disease mutations can elucidate mechanisms of disease in terms of specific interactions affected. The contents of this dissertation describe a range of research efforts I’ve led or contributed to aimed at broadening the scope of a networks-based approach to human health and disease centered around protein interaction molecular phenotypes. These efforts begin with a systematic effort to provide the resources necessary to functionally characterize rice proteins at high-throughput and direct applications to map the rice protein-protein interactome. The experimental approaches and computational analyses described here can be extended beyond rice and could provide the bases for molecular network characterization in any species. From there, I describe my contributions to validate a mutation library containing plasmid clones for over 2,000 human population and disease variants. This resource was leveraged to comprehensively measure impacts these variants on protein interaction networks to directly quantify and contextualize the extent of disruptive variants and their relationship with the human genetic background, disease, and overall fitness. Finally, I extend existing machine learning frameworks to predict protein interaction interfaces by applying protein-protein docking to construct full 3D models for these viral-host interactions between SARS-CoV-2 and human proteins. I subsequently perform mutational scanning and binding affinity calculations to predict impacts of molecular perturbations within these interactions. In doing so I explore the utility of structural interactome modelling to investigate the implications of recent evolutionary history, genetic population diversity, and potential drug repurposing on viral-host pathology through the lens of protein interactions. Cumulatively, these efforts have expanded the pace at which systematic molecular profiling of protein interaction networks can be conducted both experimentally and computationally in the Yu lab and the broader scientific community.