SAR Index: Quantifying the Nature of Structure−Activity Relationships

Abstract

Structure-activity relationships (SARs) can display very different features. Small chemical modifications of active molecules often dramatically alter biological responses. By contrast, structurally diverse molecules can have similar activity. SARs can also be heterogeneous in nature. For example, for structurally diverse molecules with similar activity, closely related analogs might have significant differences in potency. Given the inherent complexity of SARs, it has been very difficult to estimate SAR characteristics from molecular structure. On the basis of systematic correlation of 2D structural similarity and compound potency, we have developed a function termed "SAR Index" that quantitatively describes the nature of SARs and establishes different SAR categories: continuous, discontinuous, heterogeneous-relaxed, and heterogeneous-constrained. These heterogeneous SAR categories are described for the first time. Given a set of active compounds and their potency values, SAR Index calculations can estimate how likely it is to identify structurally distinct molecules having similar activity.