VMD-L Mailing List

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Wed Jul 24 2013 - 05:16:30 CDT

On Wed, Jul 24, 2013 at 11:36 AM, Mayne, Christopher G
<cmayne2_at_illinois.edu> wrote:
> Francesco,
>
> Currently ffTK does not support the parameterization of improper terms (it's on the list). It is of note; however, that there are very few cases in which improper terms are required for CGenFF (e.g. carbonyls, planar non-cyclic nitrogens). Aromatics generally do not require improper angle terms because the torsion potentials are stiff, along with the additional constraints of bonded terms within a cyclic context. It's worth taking a look through CHARMM and CGenFF parameters to see exactly there improper terms are employed and where they are not (I was surprised at how few chemical groups require them).

Hi Christopher:
Thanks a lot for your comments. I must admit that I had parm7 in mind
when I posted. I have now verified that parm7 uses impropers for
aromatics like pyridine and pyrrole, not only for their external
connections but also internally in the rings. At least as I used parm7
while waiting for the torsion calculations (some of them heavy also at
RHF/6-31G* level). But it is not my intention to confuse ideas by
comparing parm7 with charmm27.

thanks
francesco
>
> Torsion scanning does tend to be one of the more computationally intensive portions of the parameterization. There are, however, a few things that you can do to reduce the cost:
> 1) identify and reduce scanning of redundant dihedrals (due to symmetry)
> 2) increase the step size (best for freely rotatable bonds with smooth torsion profiles)
> 3) reduce the scan range (best for highly constrained torsions)
>
> Regards,
> Christopher Mayne
>
>
> On Jul 23, 2013, at 3:12 AM, Francesco Pietra wrote:
>
>> In the meantime I came across
>>
>> *****************
>> From: JC Gumbart [gumbart_at_ks.uiuc.edu]
>> Sent: 29 June 2013 13:55
>> To: Darre Castell, Leonardo
>> Cc: vmd-l_at_ks.uiuc.edu
>> Subject: Re: vmd-l: ureybradley and impropers in ffTK
>>
>> I'd like to know too! If you write some code for these functions, let us know.
>>
>> In all seriousness, it just hasn't been a priority, as they aren't
>> nearly as important or prevalent as the other terms in the force
>> field.
>>
>> On Jun 28, 2013, at 11:52 AM, Darre Castell, Leonardo wrote:
>>
>> Hi, this is probably a naive question, but I'd like to know how to
>> parametrize Urey-Bradley and Improper Torsions using ffTK tool?
>>
>> Thanks a lot in advance.
>> Leonardo
>> ********************
>>
>> May I ask what that means for the case I have prospected? That is,
>> what to do when there are aromatic rings? Impose impropers by analogy
>> to the benzene ring or do what FFTK-ScanTorsions suggests? Should the
>> latter be the choice, is it acceptable to calculate ScanTorsions at
>> RHF/6-31G*? At MP2/6-31G*, with my workstation running gamess-us
>> (workstation for MD, thus with unusable GPUs), I should work for a
>> couple of weeks 24h/24h for my complex molecule, just for
>> ScanTorsions.
>>
>> thanks
>>
>> francesco pietra
>>
>>
>> ---------- Forwarded message ----------
>> From: Francesco Pietra <chiendarret_at_gmail.com>
>> Date: Tue, Jul 23, 2013 at 9:14 AM
>> Subject: FFTK scan torsions aromatics
>> To: vmd-l_at_ks.uiuc.edu
>>
>>
>> Hope not to have missed what seems to be an obvious answer.
>>
>> FFTK ScanTorsions indicates also dihedrals involving aromatic rings,
>> such as pyridine and pyrrole, and more exotic ones, like tropylium and
>> others. Is not that a matter for impropers, forcing those rings to
>> planarity?
>>
>> thanks
>> francesco pietra
>