Jeff Henrikson (Groovescale), John Osborne (Department of Medicine, Informatics Institute, University of Alabama at Birmingham)

We present a functional pearl, signature memoization, and apply it the context of drug repurposing. To demonstrate signature memoization, we build an embedded domain specific language (DSL), Dagopoly, in Scheme and Python. We include just enough capability for computational efficiency, programmer clarity, and cross-language integration. To demonstrate application to Alzheimer’s Disease (AD) drug repurposing, we apply our DSL to transforming a list of AD genes along with a knowledge graph in order to produce a list of drug repurposing candidates. We demonstrate that random sampling of graphs dovetails nicely with signature memoization, and can dramatically reduce the time required for a typical edit/run/check cycle in data analysis exploration and development. Finally, we reflect on the DSLs fit for today’s bioinformatical needs and on the remarkable variety of related software available even when we restrict to the bioinformatical domain.