Practical ranges for protein–ligand MD, from 50 ns pose checks to microsecond-scale events, and the traces that tell you whether the run has settled.
Supervised learning needs a label on each example. Unsupervised learning looks for structure when that label is absent. The difference is in the table, before any algorithm is chosen.
ChEMBL, BindingDB, PubChem, PDBbind, and BRENDA answer different questions. The useful choice is the table that holds the measurement you mean to model.
How a ChEMBL IC50 table for one target becomes a labelled set: one row per molecule, a mean potency, and a class that leaves the uncertain middle out.
