Uncovering activity cliff generators using distribution of SALI values

Resumen

Activity cliffs are defined as compounds with high structure similarity but large potency difference. Identification of activity cliffs have a significant impact in lead optimization in medicinal chemistry, and computational applications such as the development of predictive models and the selection of queries for similarity searching. Therefore, the identification of compounds highly associated with activity cliffs in a given data set i.e., ‘activity cliff generators’, is of major relevance. Herein, we report the identification of activity cliffs and structure-activity relationships of a set of 289 synthetic compounds tested in a G protein-coupled receptor kinase, GRK. To account for information of multiple structure representations we used mean Structure-Activity Landscape Index (SALI). Structural fragments responsible for the activity are discussed.