Protein-protein interaction (PPI) modules represent the core of network medicine, leaving, for example, the "one disease-one target-one drug” dogma and KEGG pathways as a relevant representation of signaling. However, constructing these modules is debatable because most predictions and algorithms have not yet been validated. In that respect, the interactome must not be overinterpreted as a spatially realistic representation of the cell. Shortest paths, for example, may thus generate artifacts. Moreover, based on recent cell biological data, the size of PPI modules may be less arbitrary or flexible than previously assumed. Data will be presented from the "REPO4EU platform for mechanism-based drug repurposing" on how to diagnose, pharmacologically target, and clinically validate first-neighbor-based modules in disease. Finally, a module cannot be simply enriched with protein-binding drugs; it needs to be understood mechanistically, which we term a "logic module." Only then can the dysregulation be defined and diagnosed, and those drugs chosen that specifically correct the module dysregulation. Thus, to become medically relevant, “network medicine" needs to overthrow not only the current organ- and symptom-based ontology of medicine but also dogmas of bioinformatics, such as gene-KEGG pathway associations, the interactome, and how to define modules beyond PPIs