Fig. 1
Molecular chirality and fingerprints. a Correlation between chirality and heavy atom count (HAC) across ChEMBL, COCONUT, and ZINC datasets. The blue line depicts the percentage of chiral molecules relative to HAC. A steady increase in the percentage of chiral molecules is observed with increasing HAC. The yellow line represents the total count of molecules corresponding to each HAC. b Chiral shingle generation concept exemplified on a selected atom pair of polymyxin B2. The generated shingle corresponds to the pair of circular substructures (blue) separated by the shortest topological distance (red) of their central atoms. Whenever the central atom of a substructure is chiral, the atom symbol in the substructure SMILES is replaced by the Cahn-Ingold-Prelog (CIP) descriptor (R, S, r, or s), or by a question mark (?) if the stereochemistry is not defined, bracketed by two “$” characters (yellow). c Percentage of molecular shingles containing chiral information vs. percentage of chiral atoms in the molecule for MAP4C (largest diameter of four bonds). These percentages were computed using a dataset of chiral molecules uniformly sampled from the Riniker & Landrum benchmark. The high r2 and Pearson correlation coefficients underscore a strong association between the two variables