Graph based methods for drug repurposing hypothesis generation have been available since the 1980s, but historically these methods have been limited by speed of computation, availability of data, and algorithmic development. In recent years, the graph ranking community has witness a rapid growth in the development, availability, and speed of graph inference algorithms. It has now become technically feasible to perform large calculations across multiple methods. In our work we have gathered a combinatorial collection of drug-disease predictions from five academic graph groups; with numeric ranking approaches include neural network driven graph embedding, automated rule mining, and statistical graph ranking algorithms. This combinatorial collection includes results for every MONDO identifier with every possible drug, we refer to this data as All-Against-All, for all drugs against all diseases. Our resulting dataset includes 22506 diseases and 52882 chemicals with a total of 1,157,416,190 ranked drug-disease pairs with explanatory biomedical hypotheses. In our presentation, we will discuss motivations for these efforts, technical challenges and details about the implementation of the ranking computational architecture, data storage, highlight novel patterns found in the data, and discuss future plans for utilizing this knowledge to prioritize meaningful in- vivo assay work.