Pathogen-Host interaction data is core to our understanding of disease processes and their molecular/genetic bases. Facile access to such core data is particularly important for the plant sciences, where individual genetic and phenotypic observations have the added complexity of being dispersed over a wide diversity of plant species vs. the relatively fewer host species of interest to biomedical researchers. Recently, an international initiative interested in scholarly data publishing proposed that all scientific data should be "FAIR"-Findable, Accessible, Interoperable, and Reusable. In this work, we describe the process of migrating a database of notable relevance to the plant sciences-the Pathogen-Host Interaction Database (PHI-base)-to a form that conforms to each of the FAIR Principles. We discuss the technical and architectural decisions, and the migration pathway, including observations of the difficulty and/or fidelity of each step. We examine how multiple FAIR principles can be addressed simultaneously through careful design decisions, including making data FAIR for both humans and machines with minimal duplication of effort. We note how FAIR data publishing involves more than data reformatting, requiring features beyond those exhibited by most life science Semantic Web or Linked Data resources. We explore the value-added by completing this FAIR data transformation, and then test the result through integrative questions that could not easily be asked over traditional Web-based data resources. Finally, we demonstrate the utility of providing explicit and reliable access to provenance information, which we argue enhances citation rates by encouraging and facilitating transparent scholarly reuse of these valuable data holdings.