Re: [AMBER] Dihedral Parametrization in Amber19: Same QM level as for 94?

From: He, Amy via AMBER <amber.ambermd.org>
Date: Thu, 15 Feb 2024 19:06:14 +0000

Hi Jones,

I don’t know for sure how the force constant of X-CA-CA-X torsions was derived for FF94. If interested you can look at the parameter data files, for example:
$AMBERHOME/dat/leap/parm/parm94.dat
Where you can find some notes that indicate the source of the force constants.

Not all torsion parameters were individually refined and some were more of an estimate. Together, they aim to reproduce the equilibrium structure, energy, and maybe the frequencies at the equilibrium structures. However, they may not capture the true barrier of rotation or energetic properties of unphysical structures, such as distorted benzenes, when there is no data or prior knowledge to benchmark to. Especially for many improper torsions, in my understanding the values were only an estimate to suppress unphysical geometries.

(I’m sure more experienced Amber devs have better answers to this..!! )


--
Amy He
Hadad Lab @ OSU
He.1768.osu.edu
From: Jones de Andrade via AMBER <amber.ambermd.org>
Date: Thursday, February 15, 2024 at 12:03 PM
To: AMBER Mailing List <amber.ambermd.org>
Subject: Re: [AMBER] Dihedral Parametrization in Amber19: Same QM level as for 94?
Hi Dr. Simmerling. Thanks for the prompt answer (and sorry for any mistyping: using the phone since we got our optical link severed here at university).
Yes, that is what I called "traditional" protocol. My point is that this is the reported protocol for the original amber ff94/96 (only difference is that instead of fitting it would mostly focus on just the overall rotation energy barrier), and I even think it is the parametrization technique used for gaff, but it still managed no stablish the parameters of a X-CA-CA-X dihedral torsion (basically, a benzene ring).
It always bothered me how it could actually be done on that way for such aromatic atoms for the same concern you expressed: the energy would skyrocket.
So, how was the original X-CA-CA-X dihedral torsion force constant parametrized for ff94/96? Does anybody has any idea?
Thanks a lot again, and best regards,
Jones
Em 15 de fevereiro de 2024 13:39:42 BRT, Carlos Simmerling <carlos.simmerling.stonybrook.edu> escreveu:
>Hi Jones,
>not sure I'm following you. In many fitting protocols, one rotates around
>the bond of interest and calculates energy using QM or other reference
>method, and then does the same in MM, then fits parameters to minimize the
>difference. Are you asking how this is done for something like a double
>bond? the energy would be much higher, and for benzene likely impractical
>due to the coupling with bonds in the ring. but ethene can be done, and the
>barrier calculate from the 90 deg geometry.
>is that what you were asking?
>
>
>On Thu, Feb 15, 2024 at 7:58 AM Jones de Andrade via AMBER <
>amber.ambermd.org> wrote:
>
>> Hi Dr. Simmerling.
>>
>> First of all, yes, I'm sorry, I was meaning the ff19sb force field.
>>
>> And thanks for the clarification: the QM calculation level for the
>> dihedral torsions indeed changed *a lot* through time. ff19sb is proof of
>> that, and its paper kind of its testimony. ;)
>>
>> One additional question that concerns the "traditional" (lowest and
>> highest energy conformations in a torsion rotation) torsions
>> paranetrizatioms that is present at least in ff94/96, but I think still
>> works at least in gaff: while in free rotating bonds it is simple to find
>> and study those (two) conformations, on the opposite case it is sometimes
>> hard to even picture it. I mean, how are/were the maximum energy dihedral
>> torsions calculated for ethene, or even worst for a X-CA-CA-X dihedral?
>> With 90º distorted ethene and benzene molecules?
>>
>> Thanks a lot for all help,
>>
>> Best regards,
>>
>> Jones
>>
>> Em 14 de fevereiro de 2024 09:00:52 BRT, Carlos Simmerling <
>> carlos.simmerling.gmail.com> escreveu:
>> >There are a variety of force fields and also training strategies. I am not
>> >sure what you mean by "AMBER19". Amber itself is the software, and the
>> >force fields have different names. If you mean the protein ff19sb force
>> >field, I suggest reading the original article for details on the QM level
>> >for training dihedrals (which included aqueous solvation). Others, such as
>> >GAFF for small molecules, differ.
>> >
>> >On Tue, Feb 13, 2024, 10:02 PM Jones De Andrade via AMBER <
>> amber.ambermd.org>
>> >wrote:
>> >
>> >> Hi all.
>> >>
>> >> I' looking for an answer that will probably quite straightforward here.
>> >> ;)
>> >>
>> >> I'm performing the parametrization of some dihedral torsions. I know
>> >> from older amber versions and its papers (i.e, Cornell JACS 1995 and Fox
>> >> JPCB 1998) that it used to be done at MP2/6-31G*//HF/6-31G* level.
>> >>
>> >> However, since Amber94/96 a lot of evolution happened until AMBER19, so
>> >> I want to ask if there was any changes (possible improvements) on the
>> >> "protocol" concerning *especifically* dihedrals torsions
>> >> parametrization, or it it is still maintained at the same level.
>> >>
>> >> Thanks a lot in advance,
>> >>
>> >> Best regards,
>> >>
>> >> Jones
>> >> --
>> >> Jones de Andrade
>> >> (jdandrade.iq.ufrgs.br)
>> >> DFQ/IQ/UFRGS
>> >> Lattes: https://urldefense.com/v3/__http://lattes.cnpq.br/6675936210583999__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua62p5Mjyt8$<https://urldefense.com/v3/__http:/lattes.cnpq.br/6675936210583999__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua62p5Mjyt8$>
>> >> Orcid: https://urldefense.com/v3/__https://orcid.org/0000-0003-3429-8119__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua62JhFJEG4$<https://urldefense.com/v3/__https:/orcid.org/0000-0003-3429-8119__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua62JhFJEG4$>
>> >> ResearcherID: https://urldefense.com/v3/__https://publons.com/researcher/AAC-5337-2019/__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua622AyCUgI$<https://urldefense.com/v3/__https:/publons.com/researcher/AAC-5337-2019/__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua622AyCUgI$>
>> >>
>> >> _______________________________________________
>> >> AMBER mailing list
>> >> AMBER.ambermd.org
>> >> https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua628U3io3o$<https://urldefense.com/v3/__http:/lists.ambermd.org/mailman/listinfo/amber__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua628U3io3o$>
>> >>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua628U3io3o$<https://urldefense.com/v3/__http:/lists.ambermd.org/mailman/listinfo/amber__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua628U3io3o$>
>>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua628U3io3o$<https://urldefense.com/v3/__http:/lists.ambermd.org/mailman/listinfo/amber__;!!KGKeukY!wYPkBjofmsH6K2mZwPQr9-Ge6oEId6Rytj9n2aJul74SlJ1OLO0XiTVhQ8epgSCS1qSdugl-iBZAEFlgua628U3io3o$>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Feb 15 2024 - 11:30:02 PST
Custom Search