Cheminformatic modelling of β-amyloid aggregation inhibitory activity against Alzheimer's disease.

In the current research, we have developed robust two-dimensional quantitative structure-activity relationship (2D-QSAR) and pharmacophore models using a dataset of 314 heterocyclic β-amyloid aggregation inhibitors. The main purpose of this study is to determine the essential structural features which are responsible for the inhibition of β-amyloid aggregation. Prior to the development of the 2D-QSAR model, we applied a multilayered variable selection method to reduce the size of the pool of descriptors, and the final models were built by the partial least squares (PLS) regression technique. The models obtained were thoroughly analysed by applying both internal and external validation parameters. The validation metrics obtained from the analysis suggested that the developed models were significant and sufficient to predict the inhibitory activity of unknown compounds. The structural features obtained from the pharmacophore model, such as the presence of aromatic rings and hydrogen bond acceptor/donor or hydrophobic sites, are well corroborated with those of the 2D-QSAR models. Additionally, we also performed a molecular docking study to understand the molecular interactions involved in binding, and the results were then correlated with the requisite structural features obtained from the 2D-QSAR and 3D-pharmacophore models.