Abstract

Towards Understanding Phenotypic Readouts - Utilizing Multiple Types of Chemical and Biological Data to Rationalize Compound Action

Andreas Bender, Lecturer, The University of Cambridge

In our work, we explore how chemical and biological information from different domains - such as ligand-target prediction, microarray readouts, and RNA-Seq information - can be used separately, or in combination, in order to understand the mode of action of a compound (such as a hit from high-content screening) in greater detail. This process allows us to arrive at a more holistic view on compound action than either the target-based or the pathway/gene-expression based information alone. In subsequent steps these models can be used to understand phenotypic screening data, be they on the cellular or organism level. In one application on Xenopus laevis developmental screens we have analyzed the phenotypes after compound treatment with the NCI Diversity for seven different phenotypic outcomes and aimed to rationalize the biological processes involved. More recent applications involved rat studies in the area of hypnotics and the rationalization of modes of actions of mixtures of compounds using RNA-Seq data, which will be discussed in this presentation.