With the annotation of the human genome approaching completion, public-sector researchers - spurred in part by various National Institutes of Health Roadmap Initiatives - have become increasingly engaged in drug discovery and development efforts. Although large and diverse chemical libraries of 'drug-like' compounds can be readily screened to yield chemically novel scaffolds, transforming these 'chemical probes' into drugs is a daunting endeavour. A more efficient approach involves screening libraries of approved and off-patent medications; both phenotypic- and molecular target-based screening of 'old drugs' can readily yield compounds that could be immediately used in clinical trials. Using case studies, we describe how this approach has rapidly identified candidate medications suitable for clinical trials in disorders such as progressive multifocal leukoencephalopathy and amyotrophic lateral sclerosis. This approach has also led to the discovery of the molecular targets responsible for serious drug side effects, thereby allowing efficient 'counter-screening' to avoid these side effects.