FRI-530 Connecting Compound Structures and Biological Effects by Gene-Expression Profiling

Friday, October 12, 2012: 1:00 AM
Hall 4E/F (WSCC)
David Jaramillo , Computational Chemical Biology, Broad Institute, Cambridge, MA
Mathias Waver, PhD , Computational Chemical Biology, Broad Institute, Cambridge, MA
Phenotypic profiling of small molecules provides detailed insights into the cellular effects of chemical perturbations. To connect compound structures to their biological effects, we collected gene-expression profiles for more than 22,000 small molecules. The collection includes known bioactive molecules as well as novel compounds from diversity-oriented synthesis (DOS) that cover many distinct chemotypes and at the same time have defined structural and stereochemical relationships.

We identified groups of compounds with similar biological effects by clustering the gene-expression profiles. When clusters of compounds that share a pattern of gene-expression activity contain similar compounds, we can extract structure-activity relationships that report on conserved molecular skeletons, appendages, and stereocenters.

Here, we analyze a specific cluster containing both DOS compounds and known benzanilide histone deacetylase inhibitors (HDACi). All DOS compounds contain the benzanilide substructure as well as features found in another HDACi class. We relate structural similarities and differences of the compounds to their gene-expression profiles to identify features that correlate with and may modulate HDACi activity. We hold that studying the interplay between gene expression and chemical features is an effective method for suggesting novel therapeutic agents and understanding the consequences of molecular structure changes in HDACi bioactivities.