Knowledge-based prediction for alternate indications and targets for known drugs

CINF 18

A. W. Edith Chan, edith.chan@glpg.com and John P Overington. BioFocusDPI, Commonwealth House, 1 New Oxford Street, WC1A 1NU London, United Kingdom
The concept of finding new uses for known drugs represents a significantly lower risk commercial strategy compared to developing New Chemical Entities (NCEs). There are two general approaches to expanding clinical utility for a known drug: 1) predicting new indications for the compound through the known molecular target and pathway, and 2) predicting new targets (and then new indications) for a drug. Both of these approaches rely crucially on integration of multiple information sources, but rely on fundamentally different approaches for their implementation. Several of our approaches use these databases, along with a series of target sequence and compound structure similarity calculations to make predictions of likely alternate targets or bioactivities for a compound. In this presentation, we outline our approaches of building and then applying a series of highly normalized pharmacology databases to the problem of predicting the primary or alternate molecular targets for a series of known drugs. Secondly we outline the application of these databases to a series of clinical microarray datasets. Finally, some results from a large scale prediction on a collection of ‘historical' drug candidates will be shown.