Identifying and modeling structure-activity relationships (SAR) for individual molecules is a key operation in the drug development toolbox. In contrast, SAR modeling for drug combinations, where activity is defined by a synergy metric, is largely unaddressed. Here we present the first example of SAR study of synergy using the NCATS matrix screening platform. Our data indicate that SARs of synergy do exist as one might expect but is not intuitively captured by current state-of-the-art synergy metrics. Factors such as dose range, assay duration and other technical source of variations significantly confound the interpretation of synergy. To address this issue, we develop a probabilistic synergy model which quantitatively evaluates the potentiation mode between drugs, taking into account the heteroscedasticity observed in cell-based assay responses. Preliminary results suggest that this synergy metric is more robust in ranking and aggregating compound pairs with respect to mechanism of action, compared to traditional approaches such as deviations from the Bliss or Gaddum models of additivity. As a result, the approach described here can form the basis of predictive modeling of synergy SAR, enabling prospective evaluation of drug combination responses for analog series.

SSR: Structure-synergy relationships

256th National Meeting (2018)

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