Predictive methods in drug repurposing: gold mine or just a bigger haystack?

There are nearly 2000 patent and peer-reviewed literature-based examples of drug repurposing to be found at http://www.drugrepurposing.info/; yet there has recently been a spate of experimental techniques used to predict new drug repurposing opportunities. This review questions whether these new methods – from computerised modelling of drug–target interactions to retrospective analysis of clinical experience – merely add testable hypotheses without addressing their inherent validity, or whether they also partially validate the new uses so that the predictions are more likely to be successfully developed. In addition, ontological methods take existing information and link two known facts to create an unknown association. These can both enhance other methods of repurposing and provide patented, commercial products, including several case historical examples.

► There are a plethora of existing examples of new uses for existing drugs. ► However, examples are still sought, by experimental or ontological methods. ► We ask how these prospective methods can validate, or de-risk repurposing proposals.