when reading this it sounds to me like I was suggesting that the D and L
amino acids should have different charges. That really isn't true in the
Amber charge model used in the SB force fields. The charges are typically
derived for a single amino acid (often capped), and the partial charges for
the mirror image would be identical. Francois is correct that actual
charges for D and L would differ in the context of a protein, as the would
for any 2 conformations, but in our fixed-charge model they are treated the
same. I have not gone back and looked at the posts to which Francois
refers, but at least presently I do not think you need different charges
for the 2 enantiomers since they are fit using single amino acids. The only
concern should be how to build the initial conformation (as long as you are
using ff14SB at least).
The place to be careful is in the dihedral parameters, where the phases of
the cosine terms determine whether they apply equally to both enantiomers
in the mirror image conformations. As Dave Case mentioned already, the ones
we supply provide the same energy for both mirror images of the individual
amino acid.
I would strongly recommend against developing a new parameter set for the D
amino acid. First, it should not be needed, and second, unless one is very
careful you could end up with a different type of model for the standard vs
alternate chirality version, meaning you have changed not only the
chirality but the force field model as well. This makes it very difficult
to compare simulations of both and determine the underlying source of any
observed differences.
Carlos
On Wed, Apr 4, 2018 at 4:47 AM, FyD <fyd.q4md-forcefieldtools.org> wrote:
> Dear Chetna, dear Paul,
>
> As Dr Case wrote two enantiomers have the same atomic charges, although :-)
>
> If you are not aware of some 'potential' problems [2 different
> molecular orientations before MEP computation for the 2 optimized
> enantiomers (quite usual for different QM programs); different
> conformations for the 2 optimized enantiomers (quite usual error),
> bad_? accuracies for the optimized enantiomer geometries] you will end
> up with different charge sets for two enantiomers.
>
> Your best bet to derive _identical_ charges for 2 enantiomers is to
> construct one enantiomer, and then define the second one as the image
> of the first one (i.e. by replacing x by -x in a text editor).
>
> I ran a test for you using PyRED using 2-butanol (R vs S)
> (based on http://q4md-forcefieldtools.org/REDServer-Development/faq.php#20
> ):
> See
> http://q4mdfft:q4mdfft2012.cluster.q4md-forcefieldtools.org/~ucpublic1/
> ADF1ADFOmADFP3sbUADFh6PZnMFAiRXhN9I77boZa0/P3956.html
> (data available 5 days)
> The two PDB inputs:
> http://cluster.q4md-forcefieldtools.org/~ucpublic1/
> ADF1ADFOmADFP3sbUADFh6PZnMFAiRXhN9I77boZa0/P3956/Mol_red1.pdb (x, y,
> z)
> vs
> http://cluster.q4md-forcefieldtools.org/~ucpublic1/
> ADF1ADFOmADFP3sbUADFh6PZnMFAiRXhN9I77boZa0/P3956/Mol_red2.pdb (-x, y,
> z)
>
> As you can see the charge sets are identical:
> http://cluster.q4md-forcefieldtools.org/~ucpublic1/
> ADF1ADFOmADFP3sbUADFh6PZnMFAiRXhN9I77boZa0/P3956/Applet-dir_
> P3956/javascript_applet-mol2-3.html
> http://cluster.q4md-forcefieldtools.org/~ucpublic1/
> ADF1ADFOmADFP3sbUADFh6PZnMFAiRXhN9I77boZa0/P3956/Applet-dir_
> P3956/javascript_applet-mol2-4.html
> in two different FF libraries:
> http://cluster.q4md-forcefieldtools.org/~ucpublic1/
> ADF1ADFOmADFP3sbUADFh6PZnMFAiRXhN9I77boZa0/P3956/Data-R.E.D.
> Server/Mol_MM/Mol_mm1-c1.mol2
> (y)
> .<TRIPOS>ATOM
> 1 C1 -1.853784 -0.034078 0.475481 CT 1 SBT -0.3056 0.0000
> ****
> 2 H11 -2.111047 0.834071 1.074103 HC 1 SBT 0.0720 0.0000
> ****
> 3 H12 -2.632866 -0.186601 -0.262899 HC 1 SBT 0.0720 0.0000
> ****
> 4 H13 -1.839094 -0.899239 1.135382 HC 1 SBT 0.0720 0.0000
> ****
> 5 C2 -0.504655 0.147421 -0.211765 CT 1 SBT 0.3303 0.0000
> ****
> 6 H2 -0.557386 1.013189 -0.862395 H1 1 SBT 0.0675 0.0000
> ****
> 7 O2 -0.239219 -0.930352 -1.080899 OH 1 SBT -0.6980 0.0000
> ****
> 8 H22 -0.210844 -1.735175 -0.580105 HO 1 SBT 0.4313 0.0000
> ****
> 9 C3 0.635273 0.352561 0.786457 CT 1 SBT -0.0716 0.0000
> ****
> 10 H31 0.710125 -0.532232 1.418976 HC 1 SBT 0.0176 0.0000
> ****
> 11 H32 0.378075 1.177192 1.446988 HC 1 SBT 0.0176 0.0000
> ****
> 12 C4 1.982417 0.625764 0.120100 CT 1 SBT -0.0576 0.0000
> ****
> 13 H41 2.262575 -0.186352 -0.539805 HC 1 SBT 0.0175 0.0000
> ****
> 14 H42 2.765098 0.746680 0.862592 HC 1 SBT 0.0175 0.0000
> ****
> 15 H43 1.945070 1.535513 -0.472430 HC 1 SBT 0.0175 0.0000
> ****
>
> http://cluster.q4md-forcefieldtools.org/~ucpublic1/
> ADF1ADFOmADFP3sbUADFh6PZnMFAiRXhN9I77boZa0/P3956/Data-R.E.D.
> Server/Mol_MM/Mol_mm2-c1.mol2
> (-y)
> .<TRIPOS>ATOM
> 1 C1 -1.853784 0.034078 0.475481 CT 1 RBT -0.3056 0.0000
> ****
> 2 H11 -2.111047 -0.834071 1.074103 HC 1 RBT 0.0720 0.0000
> ****
> 3 H12 -2.632866 0.186601 -0.262899 HC 1 RBT 0.0720 0.0000
> ****
> 4 H13 -1.839094 0.899239 1.135382 HC 1 RBT 0.0720 0.0000
> ****
> 5 C2 -0.504655 -0.147421 -0.211765 CT 1 RBT 0.3303 0.0000
> ****
> 6 H2 -0.557386 -1.013189 -0.862395 H1 1 RBT 0.0675 0.0000
> ****
> 7 O2 -0.239219 0.930352 -1.080899 OH 1 RBT -0.6980 0.0000
> ****
> 8 H22 -0.210844 1.735175 -0.580105 HO 1 RBT 0.4313 0.0000
> ****
> 9 C3 0.635273 -0.352561 0.786457 CT 1 RBT -0.0716 0.0000
> ****
> 10 H31 0.710125 0.532232 1.418976 HC 1 RBT 0.0176 0.0000
> ****
> 11 H32 0.378075 -1.177192 1.446988 HC 1 RBT 0.0176 0.0000
> ****
> 12 C4 1.982417 -0.625764 0.120100 CT 1 RBT -0.0576 0.0000
> ****
> 13 H41 2.262575 0.186352 -0.539805 HC 1 RBT 0.0175 0.0000
> ****
> 14 H42 2.765098 -0.746680 0.862592 HC 1 RBT 0.0175 0.0000
> ****
> 15 H43 1.945070 -1.535513 -0.472430 HC 1 RBT 0.0175 0.0000
> ****
>
> However adding a D-aminoacid 'in the middle' of L-aminoacids (within
> an oligopeptide) is a different story: this is NOT an enantiomer
> related problem; this leads to diastereosiomers, which for sure share
> different atomic charge sets: this can be easily checked by deriving
> charges for a LL-diaminoacid vs a DL-diaminoacid (or better_? a
> LLL-triaminoacid vs a LDL-triaminoacid). This was discussed at least
> twice in the Amber mailing list in the past (with Dr Simmerling).
>
> I hope this helps,
> regards, Francois
>
>
> > For this, I first created an Isovaline residue (using Marwinsketch or any
> > other chemical molecule editor) and changed its chirality using Avogadro.
> > These molecules then can be used as input to the R.E.D server for charge
> > calculation and force field parameters.
> >
> > http://upjv.q4md-forcefieldtools.org/REDServer-Development/
> > go through their many tutorials on how to create such entities.
>
> > On Sat, Mar 31, 2018 at 7:05 AM, <rabi1993.iitg.ernet.in> wrote:
> >
> >> I want to work on chirality sensing of epoxides. But I have no idea
> how
> >> to create structures and force field parameters for these 'R' and 'S'
> >> enantiomers separately. Is there any tools for creating these types of
> >> structure and force field?
>
>
>
>
> F.-Y. Dupradeau
> ---
> http://q4md-forcefieldtools.org/FyD/
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Apr 04 2018 - 05:30:03 PDT
