Data-driven drug discovery (D4) exploits a comprehensive set of ‘Big’ data in order to provide an efficient path to new drug development. Particularly, more than a million bioassay dataset for drug screening became publicly available. It provides, though not complete, extensive information of bioactivity profiles for millions of compounds. First, we developed a novel systems pharmacology method to predict active compounds for a target protein, named BEAR (Active Compound Enrichment Analysis). Bear is capable of predicting active compounds of novel scaffolds and easily scalable to thousands of targets. Drug repositioning (DR) exploits polypharmacology (i.e. multiple targets for a single drug) in order to identify new indications for existing drugs. In the second part, we present the use of recent chemical genomic dataset of ~20,000 drug-induced expression profiles (LINCS). Our in silico DR method is shown to provide i) global view of transcriptional perturbation even by unknown off-targets, and ii) rich information on drug mode-of-action.
Wan Kyu Kim, Systems Biology, Ewha Womans University, Seoul, Korea
Host: Christine Vogel
Data-Driven Drug Discovery (D4)