> When an amino acid residue is modified in a pdb
> file, how can we get its prep and frcmod file (AMBER
> parameter files)?
For a _new_ residue (nucleotide, sugar, or aminoacids) _compatible_ with
existing AMBER libraries, you could do (in the order):
1- Derive new RESP (or ESP) charges for this unit. For this you can use our
R.E.D. program see: http://www.u-picardie.fr/labo/lbpd/RED/
[The RESP/ESP charges are written in a Tripos .mol2 (final output)]
You might have to use intra-molecular restraints in the RESP fit to build this
UNIT: See the FAQ http://www.u-picardie.fr/labo/lbpd/RED/FAQ-I.htm for this.
2- You need to set up the force field (FF) atom types for this unit (You can do
it by adding the FF atom types in the 6th column of this file).
3- You have to run LEaP to see which FF parameters are missing. For this you can
load the Tripos .mol2 file into AMBER8/LEaP (_instead_ of a prep.in file).
M = loadmol2 Mol_o1.mol2 # load the .mol2 file
saveoff M M.off # save the off lib. for future
saveamberparm M M.top M.crd # save the FF params
At this stage, LEaP should complain _if_ FF params are missing and list them...
* If bond or angle FF parameters are missing: For the eq. value, run
B3LYP/6-31G* geom. optimization & measure the requested value. For the FF cste
just take a value available in GAFF or parm99 for a similar bond/angle (not
that important for the global conformation); you can also execute
AMBER8/parmcal.
* If dihedral FF params are missing, it is more tricky...
You can try to take once again a dihedral available from GAFF/parm99 but I would
suggest to fit Emm(dihedral) to Eqm-Emm(non-bond) for a set of rotamers. This
procedure is quite tedious & time consumming but might be crucial since
dihedral FF params are primary determinants to the molecular conformation...
- You build your "Frcmod.New" file with these new FF params and you re-run LEaP
as follows:
xleap -f leaprc.ff99 # you use the parm99.dat FF
FRCMOD = loadamberparams Frcmod.New
loadoff M.off # you load your OFF lib. built previously
saveamberparm M M.top M.crd
It should work, I hope it helps...
Regards, Francois
F.-Y. Dupradeau
DMAG EA 3901 & Faculte de Pharmacie, Amiens, France
The Scripps Research Institute, La Jolla, CA, USA
--
http://www.u-picardie.fr/labo/lbpd/FyD.htm
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Fri Apr 08 2005 - 17:53:00 PDT
