Hi,
You should certainly define the qm_mm mask with atom subsets. Check out the
ambermask syntax.
You will need to determine which part of the chromophore you require to be
QM. If it is the whole chromophore, you have no choice but to put
link-atoms in the peptide-backbone. This is not ideal, but can be ok. I
would say that the best place is between C and CA atoms (so you'll end up
treating a little bit of the backbone QM as well). Best not to put a link
atom between C and N, as this is a more polarized bond.
--Marc
On 13 May 2014 14:09, James Starlight <jmsstarlight.gmail.com> wrote:
> Thanks Marc!
>
> Regarding bonds between atoms treated in qm and mm approximations: will it
> be more correctly to define qm_mm mask using some atom subsets (not
> entirely residue number)- e.g excluding atoms which are connected to the
> rest of the protein backbone. E.g having GFP chromophore which is
> covalently attached to the rest of the protein backbone both from its N and
> C termi- it would be logically to define connectors as chromophore's N
> (connected to the previous C atom of the adjacent amino acid) and C
> (connected to the N atom of the next a.a) atoms. So It seems that all
> chromophore atoms should be defining in the qm'mm mask besides its first C
> and last N atoms, shouldn't it? How it could be defined in amber input
> file?
> If I misunderstood It I's not well clear for me what connectors in the case
> of GFP chromophore might be placed automatically if I just define
> :name_of_Chromophore_residue as the qmmm mask.
>
>
>
>
>
> James
>
>
> 2014-05-13 16:42 GMT+04:00 Marc van der Kamp <marcvanderkamp.gmail.com>:
>
> > Some additional answers ;)
> >
> > On 13 May 2014 13:02, James Starlight <jmsstarlight.gmail.com> wrote:
> >
> > > Some additional questions:
> > >
> > >
> > > 1) Does it possible to use *.prmtop and *.inpcrd files made from the
> > > previous equilibration MD run For instance initially my ligand was
> > > parametrized by force field method and its parameters have been
> included
> > in
> > > the prmtop which I'd like to use as the input for qm_mm defining ligand
> > > treatment by ab initio? Will force field parameters of this group be
> > > ignored during qmmm?
> > >
> > > Yes. You can use the same *.prmtop and *.inpcrd (or *.rst or whatever).
> > Just select the ligand in the qmmask in the qmmm section of the
> inputfile.
> > Force field parameters will be ignored for all atoms treated QM.
> >
> >
> > >
> > > 2) How should I made *.prmtop file if I'd like to run qmmm calculations
> > > with the ab initio treatment of one of the non-standart residue which
> is
> > > covalently connected to the rest of the protein backbone (e.g as in
> > > Rhodopsin or GFP). Should I define that covalently bonds with the rest
> of
> > > the protein backbone explicitly in the conf file ?
> > >
> >
> > In the same way as a non-covalent residue, BUT you will want to think
> about
> > which atoms should be treated QM and which MM - Amber will automatically
> > place link-atoms to treat the covalent QM/MM boundary. Best is to pick
> > non-polar bonds (like CA-CB in amino-acid side-chains) as the qm/mm
> > boundary.
> >
> >
> >
> > > 3)Again I'll be very thankful for some tutorial covered that questions
> as
> > > well as some theory background of the ab initio methods implemented in
> > the
> > > amber.
> > >
> > > Others may know of tutorials (outside those on the amber website).
> > There are no 'ab initio' methods inplemented in AMBER, 'only' a range of
> > semi-empirical methods (see sqm in AmberTools manuals).
> > You can interface with QM programs to use other methods.
> > "some theory background" is a bit too vague here - there are numerous
> > textbooks that can help you, but a thorough understanding of the enormous
> > range of existing QM methods (that could be roughly characterised as
> > semi-empirical, DFT, or ab initio) will take years of study...
> >
> >
> >
> > > James
> > >
> > > --Marc
> >
> >
> > >
> > > 2014-05-13 15:39 GMT+04:00 Marc van der Kamp <marcvanderkamp.gmail.com
> >:
> > >
> > > > Yes, that is what it means. QM/MM can only be run with sander.
> > > > --Marc
> > > >
> > > >
> > > > On 13 May 2014 12:34, James Starlight <jmsstarlight.gmail.com>
> wrote:
> > > >
> > > > > Thanks Jason!
> > > > >
> > > > > Trying to launch such simulation on GPU usng pmed I've obtained
> > > > > | ERROR: ifqnt must == 0!
> > > > > | PMEMD 14 does not support QM/MM calculations.
> > > > > | Please use sander 14 instead.
> > > > >
> > > > > Input errors occurred. Terminating execution.
> > > > >
> > > > > does it mean that qm_mm calculations could not be run on GPU?
> > > > >
> > > > > James
> > > > >
> > > > >
> > > > >
> > > > > 2014-05-12 15:13 GMT+04:00 Jason Swails <jason.swails.gmail.com>:
> > > > >
> > > > > > On Mon, 2014-05-12 at 13:32 +0400, James Starlight wrote:
> > > > > > > Dear Amber users!
> > > > > > >
> > > > > > > I'd like to perform some test simulations using QM/MM method of
> > > > > > > parametrization mainly based on that tutorial
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
> https://mmtsb.org/workshops/mmtsb-ctbp_workshop_2009/Tutorials/qmmm_amber/index.htm
> > > > > > >
> > > > > > > Firstly I have just simulated water-soluble protein-ligand
> system
> > > > (this
> > > > > > > time ligand have been parametrized by gaff) and would like to
> > > perform
> > > > > the
> > > > > > > same simulation but changing treatment of my ligand to:
> > > > > > > qmmask = ':200' # number of ligand residue
> > > > > > > qmcharge = 1 # because my ligand hase total -1 charge
> > > > > >
> > > > > > Then your qmcharge should be -1, not 1.
> > > > > >
> > > > > > > qm_theory = 'PM3'
> > > > > > >
> > > > > > > as I noticed that simulation should be exactly like above
> > mentioned
> > > > > > example
> > > > > > > shouldn't it? Should I re-parametrize my ligand for such case
> > (E.g
> > > > > using
> > > > > > > REDS)?
> > > > > >
> > > > > > Parameters for a ligand you plan on treating with QM in a QM/MM
> > > > > > simulation are ignored. Residues are modeled with _either_ the
> > force
> > > > > > field parameters _or_ the semiempirical QM Hamiltonian, not both.
> > > > > >
> > > > > > HTH,
> > > > > > Jason
> > > > > >
> > > > > > --
> > > > > > Jason M. Swails
> > > > > > BioMaPS,
> > > > > > Rutgers University
> > > > > > Postdoctoral Researcher
> > > > > >
> > > > > >
> > > > > > _______________________________________________
> > > > > > 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
> >
> _______________________________________________
> 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 May 13 2014 - 07:30:04 PDT
