GDB-9 Point Group Symmetry Classification

Point group symmetry labels for the complete GDB-9 (QM9) dataset: 133,885organic molecules, each classified into one of 21 Schoenflies point groups.Labels were assigned with the pymsym library applied directly to theB3LYP/6-31G(2df,p) equilibrium geometries distributed as XYZ files in theoriginal dataset, without re-optimisation or symmetrisation. The distribution is strongly dominated by the trivial group: C1 accounts for89.67% of the molecules and Cs for 9.49%, with the remaining 19 point groupsforming a long tail in which four groups are represented by a single moleculeeach. For 157 molecules (0.117%) no stable assignment could be obtained within thegeometric tolerance of the symmetry detection procedure. These are retained andexplicitly labelled UNKNOWN rather than discarded, so that the record covers thecomplete GDB-9 set and the denominator of any downstream statistic isunambiguous. Each molecule is additionally annotated with its rotational symmetry number,atom count, electron count, molecular weight, Hill formula and a planarity flag.Column definitions, methodology and attribution to the original GDB-9 and pymsymauthors are given in the accompanying README.