A curated comprehensive dataset of asymmetric organocatalytic Mannich reaction

Asymmetric organocatalytic Mannich reactions are a cornerstone method for the stereoselective construction of C–C bonds and nitrogen-containing stereocenters. Despite the large body of literature, reaction data remain scattered across publications in heterogeneous formats, which limits their reuse for data-driven and machine-learning studies. This dataset addresses that gap. It contains 4,015 asymmetric organocatalytic Mannich reactions collected from 154 primary publications and manually curated — structures, conditions, and stereochemical outcomes were verified against the original sources rather than automatically extracted. Each entry includes the reaction components as SMILES, the catalyst and its class, reaction conditions, and the reported stereochemical outcome (ee, and where available, dr). The dataset comprises 468 unique organocatalysts, 538 unique nucleophiles, and 578 unique electrophiles (imines). Note. The number of unique organocatalyst SMILES may be smaller than 468: atropisomeric BINOL-derived catalysts can share an identical SMILES string (since standard SMILES does not encode axial chirality) while representing distinct catalysts that differ only in the axial_configuration column (R or S). Such cases are counted as separate organocatalysts. The dataset is intended as a reusable, machine-readable foundation for enantioselectivity prediction, and broader cheminformatics work on organocatalysis.