Nathan,
if I understand your explanations correctly then you simply intend to
generate parameters for a novel N-terminal residue that may be used in
conjunction with standard protein amino acid residues.
Like Carlos, I do not understand the second call of parmchk2 neither.
Just for clarification:
The definition of a force field residue within a give force field
includes atom types, atomic charges, and the topology. This definition
is read in by leap via the standard parameter files (source commands for
protein, gaff, solvent); after that specialized parameters are loaded.
When you build an actual system, e.g. via the sequence command, then
parameters for this system are collected from the available Amber
parameter files (and user provided files) according to the atom types
and the topology; the acutal parameters setup (bonds, angles, dihedrals
..) is AFAIK triggered by the saveAmberParm command in leap: if there
are parameters missing for an interaction, leap complains about that;
i.e. for all bonds in the system the corresponding pair of atom types
must be contained in the bonds parameter lists read in initially, and
the same holds for the other parameters.
Thus, the general workflow would be:
1, Create a new N-terminal residue with atom types, atomic charges, and
topology - what you already did. (Carefully check the atom types.)
2, Setup a small test sequence like { TAL GLY NME } to identify the
missing cross-residue parameters, if you combined gaff and a protein
force field, and select suitable parameters from the existing files or
generate new ones. Create an frcmod file containing the missing
parameters; then, your system has a complete force field description.
3, Setup, run, and analyze validation systems. Depending on your
estimation of the results, either go back to step 1/2 or continue with
you actual research question.
Sorry for being that general.
I think you'd just need to define the cross-terms; however, I'm not sure
about mixing ff19SB and gaff2 (see Note 1 in Amber21 manual on p.37),
but maybe others have already experience with that kind of mixing.
Best,
Anselm
Bioinformatik | NHR.FAU
Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU)
Germany
Am 21.02.2022 um 19:03 schrieb Carlos Simmerling:
> I'm not sure I understand your workflow and the goal of each step. If I
> understand correctly, you're generating parameters, then running leap to
> build the polymer, then generating parameters again? why do you need to
> generate the 2nd frcmod? if it's to make the cross terms, this doesn't seem
> like the best way since it will also try to generate parameters for the
> amino acids, and probably redo the once in your initial frcmod. maybe it's
> just not clear what you're using as input the second time - are you just
> using parm19.dat and not GAFF or anything else?
>
> As Anselm suggested, you likely want to generate parameters for the new
> residue using GAFF or something like that, and make the cross terms between
> GAFF and the protein model, and then use that frcmod in leap to build the
> system. I don't see a need for redoing the frcmod after you've already done
> it once.
> Others may be more familiar with this type of workflow and have better
> suggestions. The modified amino acid tutorial on the main Amber web site
> may also be helpful.
>
>
> On Mon, Feb 21, 2022 at 12:55 PM Nathan Black <nathanblack262.gmail.com>
> wrote:
>
>> Dr. Horn and Other AMBER Users,
>>
>> I would like to provide more context with regard to my methodology and
>> overarching research purpose. After describing my methodology in detail, I
>> will address the errors I have run into. My immediate goal is to simulate
>> interactions between polymerized micelles and organic ligands. At the
>> moment, my methodology begins with processing a .cif file corresponding to
>> the modified undecylenic acid-based tail, then adding the amino acid
>> residues in LEaP following the sequence {} command.
>>
>> The process of the input .cif file begins with the use of *antechamber* to
>> perform charge assignments using the AM1-BCC method (-c bcc) and define
>> atom types by GAFF2 (-at gaff2), giving a .ac file output. From looking at
>> past files on my research computer, the (-at amber) command was given,
>> which as I understand corresponds to PARM data types rather than GAFF2.
>>
>> Then, I used *prepgen* to omit hydrogen atoms as needed for polymerization
>> and addition of amino acid residues; the tail atom was defined (the head is
>> null) to facilitate the addition of amino acid headgroups. This gave a
>> .prepin file output.
>>
>> After this, *parmchk2 *was used to generate force field modifications for
>> the surfactant tail, using the gaff2.dat force field parameter file as an
>> input (-p $AMBERHOME/dat/leap/parm/gaff2.dat). I will designate this
>> tail-based .frcmod file as *tail.frcmod.*
>>
>> These force field modifications were loaded into LEaP, and the surfactants
>> are generated using the sequence {TAL AMR ... CAMR} command, where "TAL"
>> corresponds to the processed tail, and "AMR" corresponds to amino acid
>> residues; this command is performed after loading the leaprc.protein.ff19SB
>> force field.
>>
>> Then, *parmchk2 *is run again on each combined surfactant residue, this
>> time using the parm19.dat force field parameter file as an input. I will
>> designate the output files as *surfactant.frcmod *files.
>>
>> This concludes my process of surfactant generation in AMBER, other than
>> saving the result of the sequence {} command. As for the issues I have run
>> into, I have found that the *surfactant.frcmod *files have several "ATTN,
>> need revision" lines when following the above procedure. I have found that
>> when the same atom type and force field modification input (either GAFF2 or
>> PARM) is used for all *antechamber *and *parmchk2 *commands, there are
>> minimal "ATTN, need revision" lines.
>>
>> In my potentially naive view, shouldn't I use *gaff2 *for the
>> *antechamber* and
>> first *parmchk2 *step (as these steps correspond to the tail in the absence
>> of amino acid residues), followed by *parm19 *for the final *parmchk2 *
>> step?
>>
>> Also, how would I go about manually determining inputs for any parameters
>> that *parmchk2 *is not able to determine? Are there recommendations given
>> for this in the AMBER manual?
>>
>> Greatly appreciative,
>> NDB
>>
>> On Mon, Feb 21, 2022 at 10:31 AM Dr. Anselm Horn <anselm.horn.fau.de>
>> wrote:
>>
>>> Nathan,
>>>
>>> parameterization is IMHO not the ideal task for getting started with MD,
>>> especially if you intend to use the new ff19SB force field.
>>>
>>> In principle, one can mix parameters of different force fields;
>>> recommendations about that are given in the manuals: e.g. gaff(2) is
>>> generally considered to be well suited to work with ff14SB.
>>> However, if you use different force fields within the same residue
>>> (protein + organic), you end up with different sets of atom types, i.e.
>>> uppercase (protein) and lowercase (gaff), which requires special, mostly
>>> manually defined cross-terms with mixed atom types (e.g. torsions); the
>>> respective parameters may, however, be obtained from already existing
>>> parameter lists.
>>>
>>> Alternatively, one could use/add specialized protein atom types and just
>>> import the then missing parameters from the other force field.
>>> In your case, you had to decide whether your total system is more like a
>>> protein with some non-standard amino acid residues, or like an organic
>>> compound which has some similarity to a protein. Then you could choose
>>> your 'main' force field accordingly.
>>>
>>> In any case, you'd need to somehow validate your parameter assignment
>>> via MD simulations to see whether they work as expected.
>>>
>>> In my (limited) experience, many parameterizations or parameter
>>> assignments involve manual steps and decisions that are not easily
>>> performed automagically. It depends on your system and your research
>>> question.
>>>
>>> But others may chime in here and provide more helpful comments...
>>>
>>> All the best
>>>
>>> Anselm
>>>
>>> Bioinformatik | NHR.FAU
>>> Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU)
>>> Germany
>>>
>>>
>>> On 02/19/2022 08:04 PM, Nathan Black wrote:
>>>> Hello AMBER Users,
>>>>
>>>> I am a new graduate student that is relatively unacquainted with AMBER.
>>>>
>>>> As part of my research, I am attempting to simulate anionic surfactants
>>>> whose head(s) are standard amino acid residues, and whose tails are
>> based
>>>> upon undecylenic acid (general organic domain). As such, I am hoping
>> that
>>>> there is a way to model each surfactant such that the tail parameters
>> are
>>>> defined by GAFF2, while the head parameters are defined by ff19SB.
>>>>
>>>> It was my initial impression that the parmchk2 program could be used to
>>>> accomplish this, but the more I read through the Amber 20 manual, the
>>> less
>>>> confident I am that parmchk2 can "satisfy" missing parameters at the
>>>> surfactant head-tail boundary. It seems that parmchk2 is intended for
>> use
>>>> in molecules with nonstandard residues but will still be described by a
>>>> single force field.
>>>>
>>>> Is there any software or some other methodology that would allow me to
>>>> successfully implement this mixing of force fields without compromising
>>> the
>>>> integrity of any subsequent MD simulations? Any help/advice is much
>>>> appreciated!
>>>>
>>>> -NDB
>>>> _______________________________________________
>>>> 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
>>>
>> _______________________________________________
>> 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
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Feb 22 2022 - 04:30:02 PST
