Diverse compound sets, such as high-throughput screening (HTS) hit sets containing an unknown number of chemotypes, have traditionally been analyzed by clustering, nearest neighbors, or other scaffold-agnostic methods rather than by rigorous R-group analysis.  Here we describe how MOE SA/Report has been applied to analyze 2-4k compound subsets of the GSK screening collection having measured activity (percent inhibition or pIC50) against one or more screened targets.  Since the default scaffold auto-detection within SA/Report is tuned for datasets with many exemplars of a few scaffolds as opposed to more diverse HTS hit sets, we use an interactive scaffold selection approach. The user is allowed to pick a scaffold from the highest ranked (most frequent and largest) fragments found in the data, and frequent fragments are then found in the remaining unmapped compounds. Both these steps continue iteratively until scaffold selection is complete.  We also describe how MOE SA/Report has been integrated into project data delivery mechanisms at GSK, with reports being run automatically for several projects on their current set of compounds via custom KNIME workflows.

Dynamic SA/reports: Analyzing current project and HTS data by interactive selection of frequently-occurring scaffolds

244th National Meeting (2012)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.