Is conformation relevant for QSAR purposes? 2D Chemical representation in a 3D‐QSAR perspective

Conformation has a key role in the mechanism of interaction between small molecules and biological receptors. However, encoding this type of information in molecular descriptors for the construction of robust quantitative structure–activity relationships (QSAR) models is not an easy task and, so far, the dependence of these models on such feature has not been thoroughly investigated. In the present study, the authors explore the effects of conformational information on a 3D‐QSAR technique by comparing models built with descriptors that encode fully described tridimensional aspects (structures docked inside a biological target), with descriptors in which this information is suppressed (flat structures) or not fully described (structures with quantum‐chemically optimized geometries). As a result, the validation parameters indicate that the robustness of the models seems to be more related to the alignment aspect of the structures than to how well their tridimensional features are described. The dependence of a 3D‐quantitative structure–activity relationships (QSAR) technique on conformational information is investigated by comparing three QSAR models built with molecular descriptors containing different levels of tridimensionality description. The first model is built from 2D representations, where no conformational information is considered; a second model uses structures with optimized geometries; and, a final model employs the most likely bioactive conformations, obtained after docking the molecules into a biological target. The latter was not reliable.