3rfl5-PosePredictionProtocol.txt

Name

A hierarchical docking method, XDZ_5

Software

OMEGA 2.5/SHAFTS/MGLTools/AutoDock Vina 1.0/ITScore_v2_TF

System Preparation Parameters

maxconfs=500 #OMEGA
Gasteiger charges #Vina

System Preparation Method

All the released crystal structures of human FXR
protein-small molecule complexes were collected from the Protein Data Bank.
If there exist more than one PDB entries containing identical small molecules
binding to the same pocket, the structure with a higher resolution was kept,
resulting in a total of 26 resulting crystal structures.
Ligand conformational libraries were generated using OMEGA.
The most stable conformation found for each ligand was then used as a starting
point for docking. Preparation for docking with Vina was done using MGLTools,
in which the ligand PDB files generated with OMEGA were converted to the PDBQT
format by assigning partial charges and atom types.

Pose Prediction Parameters

Exhaustiveness=30 #exhaustiveness of global search (default=8)
Vina scoring function (empirical + knowledge-based function)
Num_modes=500 #max number of poses to generate
Re-scoring function, ITscore_v2_TF #a calibrated version of ITScore 2.0 using the SM-TF database

Pose Prediction Method

For a query compound, the SHAFTS program was employed to calculate
its structural similarities (i.e., the shape-feature similarities) with the small molecules
in the released FXR structures. If HybridScore > 1.2, the protein structure in the PDB entry
that has the best similarity score with the query ligand was used for docking. Otherwise,
ensemble docking was performed using all the 26 FXR protein structures collected from the PDB.
Then, a modified version of AutoDock Vina was used for the sampling, outputting up to 500 binding
models for further evaluation. A knowledge-based scoring function, ITScore_v2_TF, was used to
evaluate these binding models. The scoring function was originally developed with a statistical
mechanics-based iterative method using the refined set of PDBbind 2012, and was then calibrated
using 974 small molecule-transcription factor complexes from the SM-TF database with the iterative method.