AMBER Archive (2009)

Subject: Re: [AMBER] problem with parametrization

From: janzso_at_brc.hu
Date: Thu Jun 25 2009 - 04:49:50 CDT

Dear Francois,

Thanks for the heads-up, I actually already spotted that charge
problem as well, but I think I know what have I done wrong, so
hopefully I can fix it.

On the other hand, you say it won't work what I intended to do? My
plan was that in the frcmod file, I define parameters for the missing
bonds and angles like this:

BOND
C -n 490.0 1.335 (same as ff99 C-N)
c -N 490.0 1.335 (same as ff99 C-N)

ANGLE
O -C -n 80.0 122.90 (same as ff99 N -C -O)
o -c -N 80.0 122.90 (same as ff99 N -C -O)
.
.
.etc

So if I copy the parameters from parm99.dat (because these parameters
are not modified in the ff99SB force field) and define the missing
parameters, as I showed above, the calculation wouldn't work? If I
understand you correctly, you meant something like this when you
suggested that I should use Amber99SB force field parameters.

Best regards,
Gabor

Quoting FyD <fyd_at_q4md-forcefieldtools.org>:

> Dear Gabor Janzso,
>
> You want to connect your modified amino-acid* AA* to other amino-acids AAs.
> Amber99SB is used for AAs. Consequently, you have to use parm99SB
> for AA* as well.
>
> You can indeed mix GAFF and Amber99SB i.e. 2 different force fields
> ONLY when you have for instance a ligand and a protein i.e. NO
> CONNECTION between the ligand and the protein.
>
> When you have a connection "between two different parts" (in your
> case the two different parts are indentical as AA) you have to use
> the same FF.
>
> As an example:
> - Could not find bond parameter for: C - n
> => Not possible: Amber99SB + GAFF impossible because of the connection
>
> - Could not find bond parameter for: C - N
> => possible: all Amber99SB => this is what you need
> or
> Could not find bond parameter for: c - c
> => possible: all GAFF => You cannot use that in your case
>
> You have to use Amber99SB for all your peptide/protein.
> I think Antechamber has been designed (i) for organic molecules (ii)
> to be connected with GAFF. You can use GAFF in some cases; not in
> your case.
>
> Then, if I continue, your have this error:
>> ERROR: The unperturbed charge of the unit: -0.107460 is not integral.
>
> Here, there is another problem, you should have an integer value.
> This likely means your charge fit did not work or some constraints
> you used are wrong.
>
> I strongly suggest you to use R.E.D.
> http://q4md-forcefieldtools.org/RED/ where RESP inputs are
> automatically built following a well established approach. Tutorial
> are available @ http://q4md-forcefieldtools.org/Tutorial/
>
> regards, Francois
>
>
>> I am trying to create a non-standard residue with antechamber to use it
>> in tleap. This residue is the isopeptide form of the gly-ser dipeptide,
>> where the carbonyl group of the glycine is attached to the -OH group of
>> the serine sidechain. I have created and parametrized this molecule
>> following the antechamber tutorial for creating non-standard residues,
>> I also created an frcmod file with parmchk. But when I try to build a
>> polipeptide chain using my new residue, leap complains about missing
>> parameters. Basically the missed parameters are at the joining points
>> of my residue with the rest of the polipeptide.
>> The message goes like this (I built a tripeptide in this example after
>> loading leaprc.gaff and leaprc.ff99SB):
>>
>>> loadamberparams glyser.frcmod
>> Loading parameters: ./glyser.frcmod
>
> => the "glyser.frcmod" file generated by Antechamber cannot be used
> in your case.
>
>> Reading force field modification type file (frcmod)
>> Reading title:
>> remark goes here
>>> loadamberprep glyser.prepi
>> Loading Prep file: ./glyser.prepi
>>> aa = sequence { ALA GSD ALA }
>>> saveamberparm aa proba.prmtop proba.inpcrd
>> Checking Unit.
>> ERROR: The unperturbed charge of the unit: -0.107460 is not integral.
>> WARNING: The unperturbed charge of the unit: -0.107460 is not zero.
>
> => an error here. The charge fit did not work.
>
>> -- ignoring the error and warning.
>>
>> Building topology.
>> Building atom parameters.
>> Building bond parameters.
>> Could not find bond parameter for: C - n
>
> => impossible
>
>> Could not find bond parameter for: c - N
>
> => impossible
>
>> Building angle parameters.
>> Could not find angle parameter: O - C - n
>
> => etc
>
>> Could not find angle parameter: C - n - hn
>> Could not find angle parameter: C - n - c3
>> Could not find angle parameter: CT - C - n
>> Could not find angle parameter: o - c - N
>> Could not find angle parameter: c - N - H
>> Could not find angle parameter: c - N - CT
>> Could not find angle parameter: c3 - c - N
>> Building proper torsion parameters.
>> ** No torsion terms for O-C-n-hn
>> ** No torsion terms for O-C-n-c3
>> ** No torsion terms for CT-C-n-hn
>> ** No torsion terms for CT-C-n-c3
>> ** No torsion terms for o-c-N-H
>> ** No torsion terms for o-c-N-CT
>> ** No torsion terms for c3-c-N-H
>> ** No torsion terms for c3-c-N-CT
>> Building improper torsion parameters.
>> old PREP-specified impropers:
>> <GSD 2>: C3 O2 C4 O3
>> <GSD 2>: +M C6 C7 O4
>> total 5 improper torsions applied
>> 2 improper torsions in old prep form
>> Building H-Bond parameters.
>> Parameter file was not saved."
>>
>> Since I want to use the force field ff99SB for my calculations, and the
>> isopeptide residue was parametrized using gaff the missing parameters
>> are aviable at both force field, but not recognized because of the
>> capitalization. I figured I can modify the frcmod file to have leap
>> handle these angles and bonds and dihedrals, but I could not find some
>> of the dihedral parameters, neither in parm99.dat or gaff.dat:
>> O-C-n-c3
>
> => impossible
>
>> CT-C-n-hn
>
> => etc..
>
>> CT-C-n-c3 o-c-N-H o-c-N-CT c3-c-N-H Could you tell me
>> where can I find these parameters, or what parameters
>> should I use?
>
> You have to use Amber99SB force field parameters. Antechamber
> generates GAFF force field parameters.
>
>> Thanks in forward:
>> Gabor Janzso
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER_at_ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber