Key points are not available for this paper at this time.
The exploration of structure-activity relationships (SARs) of small molecules is a central aspect of medicinal chemistry. Typically, SARs are analyzed on a one-by-one basis, and chemical intuition and experience play an important role in this process. Since the 1960s, computational approaches have been developed to aid in SAR exploration that largely, but not exclusively, rely on the quantitative (Q)SAR paradigm. Accordingly, QSAR analysis has long been a mainstay of compound optimization efforts. However, the strong compound class dependence of SAR features and their intrinsic heterogeneity often pose severe constraints on the applicability of these methods. In addition to QSAR approaches, conceptually different molecular similarity methods are also applied to identify novel active compounds. In order to complement and further extend the current repertoire of computational methods, SAR analysis functions have recently been introduced that evaluate and compare SAR features on a large scale, extract SAR information from compound data sets and prioritize SARs that are promising targets for optimization. SAR analysis functions are designed to systematically profile and compare SARs contained in different data sets and characterize both global and local SAR features. Numerical SAR analysis is complemented by intuitive graphical representations of SAR landscapes.
Peltason et al. (Mon,) studied this question.