Dear Ross,
Thanks for you answer. I know that, apparently, what I am doing is not very
"physical", but I need it in order to remove some effect of Ewald in a
binding calculation I am doing. I just showed a part of it, that's why it
doesn't make much sense.
I followed your suggestion and used tleap to generate the prmtop file for
Ewald. But nothing changed. So my question is, assuming I don't have errors
in the preparation of the systems, does it sound reasonable to anybody that
using two sets of coordinate+parameter files, with the difference that one
defines a ligand alone, the other a ligand in a periodic system, but both
without water, the result of the thermodynamic integration is ~2 kcal/mol vs
40 kcal/mol? I guess that difference could be just the influence of the
periodic copies on the central copy of the ligand, but doesn't it sound too
much for an uncharged ligand?
Any suggestion will be greatly appreciated.
Ignacio
On Mon, May 23, 2011 at 4:46 PM, Ross Walker <ross.rosswalker.co.uk> wrote:
> Dear Ignacio,
>
> > I am trying to compare a thermodynamic integration done with PME vs one
> > without PME, both in vacuum. But I am obtaining strange results. Maybe
>
> This does not make any sense. The purpose of PME (and I assume by with PME
> you mean using regular Ewald , or do you mean a simple direct space
> cutoff?)
> is to give you long range electrostatics. While you can certainly build
> your
> system without solvent and put a box around it and run PME it certainly
> will
> not be 'in vacuum' since you will have all the images of your protein
> present. I think you need to think carefully here about what it is you want
> to simulate and what you actually are / can simulate.
>
> > somebody can point some error I am making inadvertently. So here is the
> > procedure I follow:
> >
> > 1) Build a system consisting only of a ligand, using xleap, and
> > generate
> > from there the coordinate and parameter files. I'll use these two files
> > to
> > run the ligand in vacuum, no PME, TI.
>
> Yes this is a vacuum simulation.
>
> > 2) Take the coordinate file (restart format) in previous step and add a
> > line
> > at the end of it, containing size and angles of the box.
>
> You should NOT do this, since you need a corresponding prmtop file that is
> also setup for periodic boundaries, not just the coordinate file. Use the
> setbox command in xleap and then save your prmtop and inpcrd file which
> will
> now be setup for periodic simulations (see caveat above about this NOT
> being
> a vacuum simulation).
>
> > 3) Take the parameter file from step 1), and add three flags,
> > SOLVENT_POINTERS, ATOMS_PER_MOLECULE and BOX_DIMENSIONS.
>
> This is hacking and will lead to a whole world of pain. Better to use the
> setbox option.
>
> Note, in both cases you probably want to set cutoff to be larger than your
> protein since otherwise you will have to deal with the fact you have erfc
> present in the PME case but not in the vacuum case. Note also that the
> result you will get with the PME case will be dependent on the size of your
> box which is this case is effectively arbitrary, hence I am not sure what
> you will actually learn of value from these calculations. It certainly
> deserves careful thinking about.
>
> All the best
> Ross
>
> /\
> \/
> |\oss Walker
>
> ---------------------------------------------------------
> | Assistant Research Professor |
> | San Diego Supercomputer Center |
> | Adjunct Assistant Professor |
> | Dept. of Chemistry and Biochemistry |
> | University of California San Diego |
> | NVIDIA Fellow |
> | http://www.rosswalker.co.uk | http://www.wmd-lab.org/ |
> | Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
> ---------------------------------------------------------
>
> Note: Electronic Mail is not secure, has no guarantee of delivery, may not
> be read every day, and should not be used for urgent or sensitive issues.
>
>
>
>
>
> _______________________________________________
> 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 24 2011 - 08:00:02 PDT
