Prediction of protonation states of ChEMBL33 compounds with MW <= 500

These are structures retrieved from ChEMBL33, which were filtered by MW <= 500, for each compounds the stable tautomer was generated by chemaxon and all stereoisomers were enumerated by RDKit. chembl_stereo.smi is a tab-separated file with SMILES for all enumerated stereoisomers, which can be used to predict protonation states by approaches using 3D representation for which stereoinformation should be important. chembl_nostereo.smi is a tab-separated file with unsteric SMILES for the same ChEMBL compounds as above and predicted protonation states at pH 7.4 by Chemaxon, MolGpKa, pkasolver, pka-predictor, Uni-pKa. MolGpKa and pkasolver were integrated in EasyDock with some fixes. chembl_nostereo_changed_only.smi is a subset of chembl_nostereo.smi with removed structures which were not changed in all programs. This may facilitate further tests and the analysis. All SMILES were generated using RDKit 2025.03.5. protonation_unipka_3runs_simple_patterns.smi - outputs of 3 independent runs of Uni-pKa for ChEMBL compounds with up to 16 enumerated stereoisomers for each. unipka_stereoisomers_inconsistent_within_run.csv - structures for which protonations states were not reproduced for the same stereoisomers