Part II · Cheminformatics

Molecules as data: representations, property prediction done honestly, graph neural networks, and structure-based design with docking. All four chapters are planned.

5 · Molecules as data soon

How a molecule becomes something a computer can compute with: SMILES strings, RDKit molecule objects, descriptors and fingerprints, and similarity search.

6 · QSAR and property prediction soon

Quantitative structure–activity relationships done honestly: scaffold splits, applicability domain, uncertainty, and the leakage traps of chemical data.

7 · Graph neural networks for molecules soon

Molecules as graphs: message passing, learned representations, and when GNNs beat fingerprints (and when they do not).

8 · Structure-based design soon

Proteins as structures: binding sites, molecular docking with AutoDock Vina, and machine-learned rescoring of poses.

See the full roadmap for how the Parts connect.