Computational Modeling of human β-secretase 1 (BACE-1) Inhibitors using Ligand Based Approaches
Multiple ligand based in silico quantitative structure activity relationship (QSAR) statistical modeling approaches and tools were used to comprehend the in vitro binding affinities (IC50) of diverse small molecule human β-secretase 1 (hBACE-1) inhibitors reported in scientific literature. Departing from the standard tradition, only 230 (~13%) small molecules were used for training the system and the prediction performance evaluated using an external validation set of 1476 (~87%) inhibitors. Overall, tangible and systematic improvements were observed as the descriptive information content and complexity of the modeling technique increased. The current results demonstrate that useful and productive models is within reach by choosing appropriate modeling techniques in spite of small datasets and diverse chemical classes, a scenario typical in HTS triaging or patent busting activities.