Run
python generate_ligands.py
The script takes all the SMILES in clinical-kinase-inhibitors.csv
and uses the OpenEye Toolkit and Schrodinger’s epik to generate a structure and a protonation state for the ligands. The resulting mol2 files are saved in the folder ligands
.
The environment variable SCHRODINGER
must point to its installation directory (on Mac OS X it is usually /opt/schrodinger/suites201X
).
This currently gives structures for all the FDA inhibitors except for Palbociclib which causes the following warnings
in function omega()
Warning: OE3DToAtomStereo is unable to perceive atom stereo from a flat geometry on atom 11 of molecule 'Palbociclib'
Warning: Palbociclib failed due to unspecified stereochemistry with strict stereo enabled
in function oequacpac.OEAssignPartialCharges()
Warning: BCC charge corrections will not be done for molecule Palbociclib
because bcc parameters for C2--C3 (C2--C3) bond can't be assigned
You can setup Yank’s simulation for different kinase structures and inhibitors for example
python setup_yank.py --receptor abl/2HYY-pdbfixer.pdb --ligand Imatinib.mol2 --solvate
and
python setup_yank.py --receptor abl/3UE4-pdbfixer.pdb --ligand Bosutinib.mol2
create a setup folder to run Yank respectively on 2HYY and imatinib in explicit solvent and 3UE4 and bosutinib in implicit solvent. The ligand will have residue name "MOL".
The script uses a simple heuristic algorithm to ensure that the receptor and the ligand are very close and do not overlap at the beginning of the simulation.