Originated by Greg Landrum (developed at Rational Discovery and maintained by the open-source RDKit community)
Core Principles & Analysis
Created by Greg Landrum and written in C++ with comprehensive Python bindings, RDKit is the standard open-source cheminformatics engine for medicinal chemistry, machine-learning QSAR pipelines, and virtual screening.
RDKit parses SMILES, SMARTS, InChI, and SDF representations to compute topological molecular fingerprints (Morgan/ECFP), Tanimoto chemical similarity, and physicochemical descriptors (LogP, TPSA, Lipinski rule-of-five counts).
Its distance-geometry conformer generator (ETKDG), maximum common substructure (MCS) matcher, and chemical reaction transforms integrate directly into Python data-science stacks and KNIME workflow nodes.