The pharmaceutical industry is experiencing unprecedented financial pressures and unsustainable research expenditures, threatening to compromise the robust development of new therapies. In response many pharmaceutical companies and translational research organizations have invested heavily in new technologies promising more efficient and productive research, including high-throughput screening, combinatorial chemistry for compound library generation, genetics and pharmacokinetics, and many others. Unfortunately these technologies have largely failed to significantly increase the efficiency of pharmaceutical research, as evidenced by the steady increase in research dollars required per new approved drug (currently estimated at nearly $2 billion). Given this paradigm, new drug discovery strategies are badly needed. In this dissertation, I describe the development and application of computational methods to drug repurposing, and evaluate the robustness of this strategy across a diverse set of human diseases. Section 1 reports the application of a range of computational techniques to drug discovery questions relevant to Alzheimer's disease (amyloid-), pain related diseases (voltage-gated sodium channels), and cancer (matrix metalloproteinase 2). In Section 2, I describe the specific application of ligand-based computational screening to inform drug repurposing efforts against Alzheimer's disease, Chagas disease, and Dengue fever. Finally, in Section 3, the importance of data quality to computationally guided repurposing is discussed in relation to the creation of curated databases for drug repurposing (the SWEETLEAD database) and for P-glycoprotein modulators (for use in model building/evaluation).