ProPKA/Obabel/VINA/AutoDock/SeeSAR Method 1
Open Babel 2.3.1/Chimera/ProPKA/MGLTools/AutoDock Vina/AutoDock 4/SeeSAR
Assumed pH 7.4
Gasteiger charges
Chimera was used to add hydrogen atoms to all FXR structures recorded in the PDB.
The protonation state was checked using ProPKA.
The binding site of each structure was considered as all residues having an atom within 5 angstrom of
the native ligand. Each structure was aligned to 1OSV (backbone alignment), and pairwise comparison
of binding sites was performed based on their RMSD. The subsequently built distance matrix was used to
perform a hierarchical clustering. We distinguished 5 main groups of binding site structures. 5 proteins, one of each group,
were then considered to run docking: 1OSV, 3BEJ, 3OKH, 3OLF and 4WVD.
Agonist, antagonists and inactives ligands from litterature were manually checked and
gathered. All molecules were all protonated at pH 7.4 using Open Babel and used to benchmark the method
and the protein that are best fitted for docking: agonists as positive control, and antagonists and inactives as negative.
A fast screening approach was performed using VINA (with both rigid and flexible binding site).
Performance of the methods and the models in discriminating actives from incatives compounds was assessed with
ROC curves, enrichment factors and predictiveness curves.
3OKH appeared to be the best model to discriminate actives from inactives molecules in both rigid and
flexible screening using VINA.
We decided to run docking of D3R ligands on the protein structure 3OKH.
To generate refined poses, we used a docking progam, AutoDock 4, rather than VINA (screening program)
with a grid spacing of 0.26A (against 1A for VINA).
AutoDock uses a genetic algorithm (GA) to generate poses. We performed 10 GA runs.
Then, SeeSAR was used to minimize and rescore the poses.
ProPKA/Obabel/VINA/AutoDock/SeeSAR Method 1
Open Babel 2.3.1/Chimera/ProPKA/MGLTools/AutoDock Vina/AutoDock 4/SeeSAR
Assumed pH 7.4
Gasteiger charges
Water molecules and other heteratoms were removed
Protein residue protonation considering local environment
Chimera was used to add hydrogen atoms to all FXR structures recorded in the PDB.
The protonation state was checked using ProPKA.
The binding site of each structure was considered as all residues having an atom within 5 angstrom of
the native ligand. Each structure was aligned to 1OSV (backbone alignment), and pairwise comparison
of binding sites was performed based on their RMSD. The subsequently built distance matrix was used to
perform a hierarchical clustering. We distinguished 5 main groups of binding site structures. 5 proteins, one of each group,
were then considered to run docking: 1OSV, 3BEJ, 3OKH, 3OLF and 4WVD.
Agonist, antagonists and inactives ligands from litterature were manually checked and
gathered. All molecules were all protonated at pH 7.4 using Open Babel and used to benchmark the method
and the protein that are best fitted for docking: agonists as positive control, and antagonists and inactives as negative.
A fast screening approach was performed using VINA (with both rigid and flexible binding site).
Performance of the methods and the models in discriminating actives from incatives compounds was assessed with
ROC curves, enrichment factors and predictiveness curves.
3OKH appeared to be the best model to discriminate actives from inactives molecules in both rigid and
flexible screening using VINA.
We decided to run docking of D3R ligands on the protein structure 3OKH.
To generate refined poses, we used a docking progam, AutoDock 4, rather than VINA (screening program)
with a grid spacing of 0.26A (against 1A for VINA).
AutoDock uses a genetic algorithm (GA) to generate poses. We performed 10 GA runs.
Then, SeeSAR was used to minimize and rescore the poses.
Genetic algorithm
Exhaustiveness=10
SeeSAR minimization
HYDE scoring function (HYDE estimates binding free energy based on two terms for dehydration and hydrogen bonding only)
Docking runs were executed with the above specified parameters while default parameters
were applied for the rest of the variables. All 10 poses returned by AutoDock were submitted to SeeSAR minimization
and rescoring (using HYDE scoring function). Poses were selected based on the best HYDE score. Each pose was
checked manually, abberant poses were discarded, and lower score poses were selected instead.
Two ligands (FXR_5 and FXR_33) displayed bad scores while they were very similar to ligands co-crystallized in FXR structures
(3FLI and 3FXV, respectively). We build them up from the 3FLI ad 3FXV co-crystallized ligands. We used the same protocol (SeeSAR)
to minimize and score the FXR_5/3FLI and FXR_33/3FXV complexes.