Hello,
My suggestion is to take each separate small molecule, parameterize it using
antechamber if that works for you (maybe use R.E.D. if you need more
stringent charge derivation), then load it into leap and use the "saveoff"
procedure to save an amber library file. For example, suppose you have 2
small molecules molecule1.pdb and molecule2.pdb. Use R.E.D. or antechamber
as you will, and get the files molecule1.mol2 and molecule2.mol2 with
partial charges for each atom. Then load them into leap as such:
mol1 = loadmol2 molecule1.mol2
mol2 = loadmol2 molecule2.mol2
saveoff mol1 Molecules.lib
saveoff mol2 Molecules.lib
This will create a library file "Molecules.lib" that has all of the bonding
and charge information for each of your molecules. Then you can use packmol
to create your combined PDB file, which should have the molecules in
Molecules.lib each with the same residue name and atom names. Then, you can
load this library into leap:
loadoff Molecules.lib
total_mol = loadPDB packmol_structure.pdb
And then you just have to make sure that you have parameters for all of the
atom types and bonds/angles/dihedrals. You can use parmchk on each of the
original mol2 files to get these.
Hope this helps,
Jason
On Fri, Sep 24, 2010 at 9:02 AM, Bozell, Joseph John <jbozell.utk.edu>wrote:
> Thanks very much for the lead on Packmol...nice little program. I generated
> a pdb file that contained a pair of the molecules I want to work with, and
> it would be a simple matter to generate a multi-molecular array, solvated
> or
> unsolvated, in the same manner.
>
> However, if I can request a little more handholding, it appears that I am
> still at the same point that I was earlier. The resulting packmol file (or
> any file containing 5, 15, 100 identical molecules that might be generated
> in Packmol) is effectively a pair of ³non-standard residues², and as
> implied
> in the manual and tutorials, needs to have the force field parameters and
> structural information that are generated in antechamber, right? Yet
> antechamber, according to the info on the mail archives, is generally only
> useful up to around 100 atoms (this pair would have 132 atoms), and per
> David Case's earlier reply, only intended for operation on a single
> molecule
> (and it works beautifully under those conditions, BTW). Nonetheless, when I
> enter either the pdb file containing the molecular pair generated by
> packmol
> or an earlier pdb file of the pair generated in Gaussian into antechamber,
> everything runs as expected and a prepin file is generated:
>
> antechamber i abxylalpair.pdb fi pdb o abxylalpair.prepin fo prepi c
> bcc s 2
>
> But when parmchk is run on the prepi file:
>
> parmchk i abxylalpair.prepin f prepi o abxylal.frcmod
>
> I receive a ³segmentation fault² error message immediately.
>
> Am I missing the point of Packmol? Does the pdb file that it generates
> allow
> you to go directly to LEaP and generate the coordinate and topology files
> for sander? It would seem not, since an attempt to enter the molecular file
> (with a TER card inserted between the two molecules/"residues") into LEaP:
>
> abxylal = loadpdb ³abxylalpair.pdb²
> edit abxylal
>
> LEaP creates all 132 atoms leading to a representation with small diamonds
> for the atoms but no bonds (connectivity) between them. This isn¹t
> surprising with two nonstandard residues. It would seem that the most
> direct
> approach would still be to generate all atomic coordinates and charges in
> antechamber on a single molecule, and then create several clones of the
> original at different coordinates in space so that it could be interpreted
> correctly in LEaP. But am I asking for a feature not available in the
> software?
>
> Sorry for the continuing questionsŠthe Amber family of programs runs
> smoothly for a single simple molecule, and I am trying to fill in
> substantial gaps in my understanding of program operation so that I can
> build more complex systems.
>
> Joe Bozell
> University of Tennessee
>
> On 9/23/10 10:13 AM, "M. L. Dodson" <activesitedynamics.comcast.net>
> wrote:
>
> > See below.
> > On Sep 23, 2010, at 8:46 AM, Bozell, Joseph John wrote:
> >
> >> Thanks for the initial feedback...I guess I'm still not clear on the
> steps
> >> necessary to execute the simulation I want. By following the manual and
> >> various online tutorials, I am able to successfully carry out
> antechamber on
> >> a **single** copy of my molecule to generate charges, and can then
> minimize
> >> the molecule, solvate it, and generate an MD simulation and
> movie...quite
> >> straightforward and effective.
> >>
> >> However, I now want to generate an appropriate set of files,
> coordinates,
> >> charges, etc. that contain 2 (or more) copies of this same molecule,
> solvate
> >> them, and carry out an MD simulation to see how/if they interact as a
> probe
> >> of the first steps in self assembly. Is there a simple way to take the
> >> optimized structure generated in amber for a single molecule and just
> clone
> >> multiple copies of it for evaluation in an MD run? Xleap does not appear
> to
> >> have a way to carry out a molecular copy step, or the ability to arrange
> an
> >> array of molecules in various conformations. Sirius 1.2 allows me to
> import
> >> a couple of molecules and position them independently, but does not seem
> to
> >> be able to save the resulting molecular pair as a single file of
> coordinates
> >> (maybe I haven't dug deeply enough into its operation??). The closest
> >> analogies I've seen involve docking a substrate in an enzyme, but I have
> >> been unable to translate those methodologies to the simpler molecules
> I'm
> >> working with.
> >>
> >> Again, thanks for any guidance,
> >>
> >> Joe Bozell
> >> University of Tennessee
> >
> > There are ways of getting LEaP to do what you want, I believe, but the
> packmol
> > program (http://www.ime.unicamp.br/~martinez/packmol/<http://www.ime.unicamp.br/%7Emartinez/packmol/>)
> was created to do
> > exactly what you want. Look at the example of a mixed urea-water box,
> IIRC.
> >
> > Of course you will need the concentration you want to simulate -> the
> number
> > of
> > molecules of your solute in the simulation cell. Plus the number of
> waters to
> > achieve a density of ~1.0. My experience is that simulation cells
> created
> > with
> > packmol drop precipitously in energy in the first few steps of dynamics
> due to
> > reorientation of the molecules leading to a lower electrostatic term, so
> > an initial short dynamics run at 100K is a good (but not strictly
> necessary)
> > idea. Then proceed on with the regular equilibration of temperature,
> density,
> > etc.
> >
> > Since simulation systems set up with packmol start out containing
> solvent, if
> > you need a vacuum prmtop (e.g., for postprocessing analysis), you will
> have to
> > strip the solvent, then generate the prmtop on the stripped PDB. Kind of
> > backwards to the ordinary way of doing things.
> >
> > Hope this helps,
> > Bud Dodson
> >
> >> ------ Forwarded Message
> >> From: Joe Bozell <jbozell.utk.edu>
> >> Date: Wed, 22 Sep 2010 14:20:36 -0400
> >> To: "amber.ambermd.org" <amber.ambermd.org>
> >> Subject: MD on an array of small molecules
> >>
> >> I am attempting to use Amber for MD runs on a collection of small
> molecules
> >> to observe their interaction during a self assembly process. I am able
> to
> >> carry out MD runs on a single molecule in a water shell using
> antechamber
> >> and subsequent steps. However, how does one carry out the same process
> on
> >> multiple copies of a single molecule within a solvent box?
> >>
> >> I have a pdb file for a pair of these molecules, and have used
> antechamber
> >> to generate a prepi file. However, parmchk on the resulting file gives a
> >> "segmentation fault" message, suggesting that the number of atoms may be
> too
> >> large.
> >>
> >> Is there a general method for carrying out this type of experiment
> within
> >> Amber?
> >>
> >> Thanks,
> >>
> >> Joe Bozell
> >> University of Tennessee
> >>
> >>
> >> ------ End of Forwarded Message
> >>
> >>
> >> _______________________________________________
> >> 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
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Graduate Student
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Sep 24 2010 - 06:30:09 PDT
