[gmx-users] Solvatebox from itp GAFF organic solvents and amber solute.

Justin Lemkul jalemkul at vt.edu
Wed Aug 5 23:40:33 CEST 2015

On 8/5/15 4:32 PM, Sharapa wrote:
> Perhaps my explanation is very confusing, while i am trying to solve at least
> three problems at the same time.
> 1) I would like to use equilibrated solvent from virtualchemistry, not to get
> more errors by equilibrating attempts. Seems i can use "solvate" similar to
> http://blog.sciencenet.cn/blog-365047-808554.html (where was genbox as previous
> version) (i do not need mixing solvents, i saw your hexane-water tutorial, i am
> trying to understand them all at the same time). Anyway attempts to give to
> "solvate" a itp topology, like it was done in chinease manual for top file for
> genbox fails. Seems now its separated to "include", I'll read manual about that.

Well, I can't read Chinese, so I have no idea what the tutorial is having you 
do, but the basic principles are:

1. gmx solvate deals exclusively with the coordinates, -cp for the solute and 
-cs for the solvent

and since solvate only automatically updates the topology in the case of water:

2. Add #include statement for solvent topology manually
3. Update the number of molecules of solvent in [system] manually

> 2) Single molecule - my solute. For which i have gaff topologies generated in
> Amber. I hope to use amb2gmx.pl to get sometopology that i can "include" after
> "insert-molecules" of my solute into the box of solvent that i describe in point 1.

It's even easier than using insert-molecules (my understanding was that you had 
a co-solvent situation, which apparently is not the case):

gmx solvate -cp solute.pdb -cs solvent.pdb

This is no different than standard tutorial workflows of solvating a protein.

> 3) Usage of GAFF in a simulation. This i will try to understand if i will manage
> to prepare box with solute and solvent and some sencefull topologies.

When dealing with non-standard systems, you really have to be familiar with the 
force field organization and all file formats.  A topology (.top) is just a 
listing of the molecules in the system, and can mention parameters specifically 
or can #include them (that just means copy and paste from another file).  For 
something like what you're talking about, presumably the following skeletal .top 
is basically what you need:

#include (some parent AMBER FF)
#include "solute.itp"
#include "solvent.itp"

[ system ]

[ molecules ]
solute  1
solvent N

where "solute" and "solvent" in [molecules] are the actual [moleculetype] names.


> Yes, i understand that for you its probably just usage of few trivial commands,
> but i am more used to another part of computational chemistry and so syntax and
> logic of Amber and Gromacs are very new, specific and confusing for me.
> Thanks a lot for your advices.
> On 05.08.2015 22:11, Justin Lemkul wrote:
>> On 8/5/15 2:59 PM, Sharapa wrote:
>>> Well, then i am confused again. "Liquid structure" pdb-s on virtualchemistry -
>>> are they of any use directly in creating solvatebox, or they just shows what
>>> system was equilibrated and topological parameters taken from?
>> A pre-equilibrated box is just that - a liquid structure suitable for further
>> simulation or use as a solvent in another system.  As for "creating
>> solvatebox," I have no clue what you're talking about here. Do you mean simply
>> using that liquid as a solvent? That's straightforward use of gmx solvate.
>> But you've alternately referred to single molecules, AMBER topologies, GROMACS
>> .itp files, and liquid coordinate files, so I honestly have no idea what the
>> primary question is any more.
>> -Justin
>>> On 05.08.2015 20:49, Justin Lemkul wrote:
>>>> On 8/5/15 2:31 PM, Sharapa wrote:
>>>>> Thanks a lot, that was very usefull.
>>>>> I will try to collected all collected theory in practice.
>>>>> On 05.08.2015 20:08, Justin Lemkul wrote:
>>>>>> gmx insert-molecules -ci single_molecule.pdb -o box.pdb -box X -nmol N
>>>>>> will insert N molecules into a cubic box of X-nm edge length. That is for
>>>>>> pure
>>>>>> solvent, and it's the same as the first step in creating a mixed solvent
>>>>>> (which is then followed by another round of gmx insert-molecules to add the
>>>>>> co-solvent).
>>>>> Single_molecule - am i right that i should use pdb of liquid structure
>>>>> (that is
>>>>> a rather big cube of solvent with thousands of atoms in it)
>>>> My instructions were for how to create a starting liquid structure (which will
>>>> require LOTS of equilibration) from a single molecule. Hence
>>>> "single_molecule.pdb" as its name, and "box.pdb" is the full liquid structure
>>>> with multiple molecules and however many atoms that winds up being.
>>>> -Justin
>>>>>>> Taking topologies from amber seems to be more trivial procedure, with
>>>>>>> amb2gmx.pl, but is it independent from forcefield. Dont i have to specify in
>>>>>>> calculation GAFF ?, and if yes - i dont see it as typical Gromacs
>>>>>>> forcefield,
>>>>>>> should i make smth extra for enabling GAFF?
>>>>>> That depends on what the .itp files contain.  If they have a [defaults]
>>>>>> directive, all the force field parameters, etc. then they're self-contained
>>>>>> and can be used as-is, then you're done. If they require parameters from some
>>>>>> parent force field (more likely, since you're dealing with frcmod files),
>>>>>> then
>>>>>> you need to utilize them in the context of whatever parent force field you
>>>>>> were using in AMBER:
>>>>>> #include (some AMBER force field)
>>>>>> #include (whatever your solvent .itp is)
>>>>>> etc.
>>>>>> -Justin
>>>>>>> Thanks in advance,
>>>>>>> Dmitriy
>>>>>>> On 05.08.2015 19:38, Justin Lemkul wrote:
>>>>>>>> On 8/5/15 11:23 AM, Sharapa wrote:
>>>>>>>>> Dear Justin,
>>>>>>>>> Thanks a lot for your links.
>>>>>>>>> I looked them once again (i googled them for few times while i am stucked
>>>>>>>>> with
>>>>>>>>> this problem) and still do not understand how they can help me.
>>>>>>>>> I will try to go throw manual once again.
>>>>>>>> I don't understand what the problem is.  You say you have topologies and
>>>>>>>> coordinates, so what's missing?  You have coordinates for a single
>>>>>>>> molecule,
>>>>>>>> but not a full box?  That's trivial to construct with gmx
>>>>>>>> insert-molecules or
>>>>>>>> genconf -nbox. All of that is described in those links.
>>>>>>>> -Justin
>>>>>>>>> Best regards,
>>>>>>>>> Dmitriy
>>>>>>>>> On 05.08.2015 01:20, Justin Lemkul wrote:
>>>>>>>>>> On 8/4/15 12:09 PM, Sharapa wrote:
>>>>>>>>>>> Thanks a lot for fast answer.
>>>>>>>>>>> Yes, my fault, i didnt specify - i have coordinates for my molecule
>>>>>>>>>>> (from
>>>>>>>>>>> dft
>>>>>>>>>>> optimisation or from any other minimisation). And i have coordinates for
>>>>>>>>>>> equilibrated solvents from virtualchemistry. So i have coordinates,
>>>>>>>>>>> amber
>>>>>>>>>>> topologies for solute and itp for solvent. And no good ideas on
>>>>>>>>>>> combining
>>>>>>>>>>> them
>>>>>>>>>>> together in one solvatebox.
>>>>>>>>>> http://www.gromacs.org/Documentation/How-tos/Mixed_Solvents
>>>>>>>>>> http://www.gromacs.org/Documentation/How-tos/Non-Water_Solvation
>>>>>>>>>> -Justin
>>>>>>>>>>> On 04.08.2015 17:26, Justin Lemkul wrote:
>>>>>>>>>>>> On 8/4/15 10:45 AM, Sharapa wrote:
>>>>>>>>>>>>> Dear users and developers of Gromacs,
>>>>>>>>>>>>> I am new in MD in general and in Gromacs particulary.
>>>>>>>>>>>>> I want to make MD of small organic molecule in few organic solvents.
>>>>>>>>>>>>> My molecule containes Si and I made GAFF parameters for it. I would
>>>>>>>>>>>>> like to
>>>>>>>>>>>>> use
>>>>>>>>>>>>> GAFF in my simulation, expecting it to be not too bad for
>>>>>>>>>>>>> describing my
>>>>>>>>>>>>> system
>>>>>>>>>>>>> (that also contain pi-pi stacking).
>>>>>>>>>>>>> So far using ambertools I have .frchmod and prmtop files for my
>>>>>>>>>>>>> solute. I
>>>>>>>>>>>>> found
>>>>>>>>>>>>> .itp GAFF topology and .pdb GAFF Liquid structure files for solvents
>>>>>>>>>>>>> that
>>>>>>>>>>>>> i am
>>>>>>>>>>>>> interested in on http://virtualchemistry.org/ site. Nevetherless i
>>>>>>>>>>>>> still
>>>>>>>>>>>>> dont
>>>>>>>>>>>>> see the way how to create solvatebox with solute starting from data
>>>>>>>>>>>>> described
>>>>>>>>>>>>> above. I didn't found a tutorial describing similar procedures. If
>>>>>>>>>>>>> anyone
>>>>>>>>>>>>> can
>>>>>>>>>>>>> recommend me anything - i would be extremely thankful.
>>>>>>>>>>>> Because coordinates can't be built from topologies. See
>>>>>>>>>>>> http://www.gromacs.org/Documentation/File_Formats/Coordinate_File#Sources
>>>>>>>>>>>> for
>>>>>>>>>>>> ideas.
>>>>>>>>>>>> -Justin


Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441


More information about the gromacs.org_gmx-users mailing list