I have a dataset of ligands ( from BindingDb databases) that target my protein of interest. For my 3D QSAR study, I wish to classify them into congeneric series of compounds. How can I approach it using Schrödinger Software? And what is the ideal number of ligands for build up my test model?