Hi all,
I am trying to provide the force constants and distance/angles in frcmod.user for a SECOND (follow-up) submission to R.E.D server so I have a complete amber force field.
Unfortunately I have no experimental detail about the structure, all I have is the ab initio optimised structure from the original (first) R.E.D. run. Thankfully, there are only five bonds which are unknown to ff99sb, but that is around 20 dihedral angles.
For the bond length and angles I simply measured the bond length and angles from the ab initio structure, and found the nearest matching set of element bond pair or angle triplets and used the force constant (V kcal...) for those. This makes sense, providing the atom element is the same then the force constant is relative to the distance/angle.
Unfortunately I do not know how to match dihedral angles to known dihedrals. In the ff99sb force field every dihedral angle is either a 0 or 180. This is really not the cae with my ab initio structure. Any suggestions regarding making a good approximation of the force constants from the force field dihedral angles would be great.
Thanks
Anthony
Anthony,
What type of molecule are you trying to produce? I am not super familiar with RED. However, if you are trying to produce the parameters for a small ligand molecule, then you could try using a local installation of Ambertools (if you have access to one - academic use should be free). As long as you have some sort of optimised geometry in a compatible format (xyz,mol2,Gaussian output, etc.), you can make use of the array of programs contained within "antechamber." While ff99sb may not have all the bonds and dihedrals that you are looking for, the GAFF forcefield might (depending on the atoms of interest). The small amount of manual labor involved would be in changing the atom types in your molecule to the GAFF-compatible lower-case types. After that, producing the final frcmod will come from two steps:
1) Run the file containing your optimised charge and geometry (molecule.mol2, for example) through antechamber to get an AmberPrep file:
antechamber -fi mol2 -f molecule.mol2 -fo prepi -o molecule.prep -nc (integer charge)
2) Run the prep file through parmchk2 to get the missing dihedral and bond parameters:
parmchk2 -f prepi -i molecule.prep -o frcmod.molecule -a Y -w Y
Using parmchk2 in this manner will reference your molecule against the general forcefield, and it should hopefully pull the values you need into the frcmod. Now, if you want to build a system containing "molecule", then you need to include the following commands in a tleap script:
source leaprc.ff99SB
source leaprc.gaff
loadAmberPrep molecule.prep
loadamberparams frcmod.molecule
After that, you should be ready to roll!
Anthony,
What type of molecule are you trying to produce? I am not super familiar with RED. However, if you are trying to produce the parameters for a small ligand molecule, then you could try using a local installation of Ambertools (if you have access to one - academic use should be free). As long as you have some sort of optimised geometry in a compatible format (xyz,mol2,Gaussian output, etc.), you can make use of the array of programs contained within "antechamber." While ff99sb may not have all the bonds and dihedrals that you are looking for, the GAFF forcefield might (depending on the atoms of interest). The small amount of manual labor involved would be in changing the atom types in your molecule to the GAFF-compatible lower-case types. After that, producing the final frcmod will come from two steps:
1) Run the file containing your optimised charge and geometry (molecule.mol2, for example) through antechamber to get an AmberPrep file:
antechamber -fi mol2 -f molecule.mol2 -fo prepi -o molecule.prep -nc (integer charge)
2) Run the prep file through parmchk2 to get the missing dihedral and bond parameters:
parmchk2 -f prepi -i molecule.prep -o frcmod.molecule -a Y -w Y
Using parmchk2 in this manner will reference your molecule against the general forcefield, and it should hopefully pull the values you need into the frcmod. Now, if you want to build a system containing "molecule", then you need to include the following commands in a tleap script:
source leaprc.ff99SB
source leaprc.gaff
loadAmberPrep molecule.prep
loadamberparams frcmod.molecule
After that, you should be ready to roll!
Hi Samuel!
What would be the difference if I would skip the first step and run parmchk2 for the mol2 file itself?
Thanks in advance
- Badges
- Science topic