Data for Umol

posebusters_benchmark_set.tar.zst - files for the prediction (features to Umol/Umol-pocket) and scoring of the pose busters benchmark for Umol/Umol-pocket 

posebusters_pred_native.tar.zst - pdb and sdf files of proteins and ligands. Includes native structures, predicted structures and relaxed predicted structures with plDDT in the B factor column. Refers to Umol-pocket.

posebusters_pred_structures_no_pocket.tar.zst - pdb and sdf files of proteins and ligands. Includes predicted structures. Refers to Umol without pocket information.

posebusters_scores.csv - contains ligand RMSD and other metrics for the unrelaxed structures predicted with Umol-pocket (40000 steps).

posebusters_no_pocket_unrelaxed.csv - contains ligand RMSD and other metrics for the unrelaxed structures predicted with Umol (60000 steps).

binding_db_results.csv - results for binding db predicted with Umol.

binding_db_above_85.tar.zst - structures and results from binding db with ligand plDDT>85

params40000.npy - the parameters used in Umol-pocket (best check point)

valid_meta.csv - ids and meta for validation set complexes

valid_scores.tar.zst - scores (ligand RMSD) on the validation set for both Umol and Umol-pocket

calibration_meta.csv - structural clusters for affinity set

calibration_set_affinities.csv - affinity values (only Kd <1000 nM was used in the paper)

train_losses.csv - losses for training Umol-pocket

train_meta.csv - ids and meta for train set complexes

params60000.npy - the parameters used in Umol (best check point)

cross_dataset_tanimoto_similarity.csv - Tanimoto similarity of training set to PoseBusters test set

af_pred.tar.zst - AlphaFold2 predictions for the PoseBusters test set used with DiffDock.

diffdock_af_pred.tar.zst  - pdb and sdf files of proteins and ligands for DiffDock+AF on the PoseBusters test set. Also includes a csv with l-RMSD scores.

neuralplexer.tar.zst  - pdb and sdf files of proteins and ligands for NeuralPlexer on the PoseBusters test set. Also includes a csv with l-RMSD scores.

rfaa.tar.zst - pdb  files of proteins and ligands for RFAA w/o template information on the PoseBusters test set. Also includes a csv with l-RMSD scores

hhblits.tar.zst - PoseBusters hhblits MSAs used to predict the AF, Umol and RFAA protein structures.