This PR adds newly parametrized kinase inhibitor files, under the folder
epik_inhibitors/output-1Aug2018 .
Be aware: the script used has been updated to use python 3.5+. The original script was limited to <=2.7.
Additionally, the script now properly checks for any failures during critical parts of the procedure,
that were simply ignored without warning in the old version. Now, every failed molecule
will be highlighted in the output log, and the protonation states that fail charging will be stored
under a Failed_Molecules folder for inspection.
Errors were encountered for Erlotinib state 2, Vemurafenib state 3, and Dasatinib state 6.
Example:
Warning: OEMMFFParams::PrepMol() : unable to type atom 0 C
Warning: OEMMFFParams::PrepMol() : unable to type atom 33 H
Warning: force field setup failed due to missing parameters for molecule ~{N}-(2-chloro-6-methyl-phenyl)-2-[[4-[4-(2-hydroxyethyl)piperazin-1-yl]-2-methyl-1,2-dihydropyrimidin-6-yl]amino]-3,4-dihydrothiazol-2-ylium-5-carboxamide
omega returned error code 0
These were all errors produced by Openeye charge assignment, stating that MMFF atom types could not be assigned to one or two atoms in the molecule. This assignment I believe is inferred on the geometry and bond order, though I could be mistaken.
Typically, minimizing with MMFF inside another tool does the trick. For Erlotinib, Vemurafenib and Dasatinib, all protonation states produced by epik were minimized with mmff using the maestro Forcefield feature. After doing so, errors were no longer encountered for Erlotinib state 2, Vemurafenib state 3. These files were saved with the mmff-minimized.mol2 suffix for reproducibility. Default settings used, with the exception of the forcefield, which was set to mmff. See here:
Additionally, state 6 of Dasatinib somehow ended up with a strange bond order when saved as mol2 by any schrodinger tool I tried.
I updated the bond table in the mol2 file manually, since I could not find a clear programmatic way to do so. New bond order as follows:
Calculation now succeeds. File is saved with the -bond-fix.mol2 suffix
This PR adds newly parametrized kinase inhibitor files, under the folder
epik_inhibitors/output-1Aug2018.Be aware: the script used has been updated to use python 3.5+. The original script was limited to <=2.7.
Additionally, the script now properly checks for any failures during critical parts of the procedure, that were simply ignored without warning in the old version. Now, every failed molecule will be highlighted in the output log, and the protonation states that fail charging will be stored under a
Failed_Moleculesfolder for inspection.Errors were encountered for Erlotinib state 2, Vemurafenib state 3, and Dasatinib state 6.
Example:
These were all errors produced by Openeye charge assignment, stating that MMFF atom types could not be assigned to one or two atoms in the molecule. This assignment I believe is inferred on the geometry and bond order, though I could be mistaken.
Typically, minimizing with MMFF inside another tool does the trick. For Erlotinib, Vemurafenib and Dasatinib, all protonation states produced by epik were minimized with mmff using the maestro
Forcefieldfeature. After doing so, errors were no longer encountered for Erlotinib state 2, Vemurafenib state 3. These files were saved with themmff-minimized.mol2suffix for reproducibility. Default settings used, with the exception of the forcefield, which was set to mmff. See here:Additionally, state 6 of Dasatinib somehow ended up with a strange bond order when saved as mol2 by any schrodinger tool I tried.
I updated the bond table in the mol2 file manually, since I could not find a clear programmatic way to do so. New bond order as follows:
Calculation now succeeds. File is saved with the
-bond-fix.mol2suffix