[gmx-users] Re: genion and a few other problems

[Justin A. Lemkul]

Quoting Manik Mayur <manik.mayur at gmail.com>:


> > Exact command lines are much more useful than your interpretation of what
> > you
> > typed.  If you just used genion -nn, that's probably insufficient
> > information.
> >
>
> Sorry for the confusion but genion won't add "Cl-" on its own unless I
> provide it with the relevant options (providing which I thought would be
> impertinent here). The only concern here was to find out "why Cl was listed
> in the [ molecules ] section of the topol.top file", hence the error as no
> such molecule was defined.
>
> But I found a workaround as to generate a .pdb file using genion, and then
> use pdb2gmx to generate .gro and topol.top files, which I see automatically
> generates a .itp file corresponding to "Cl-" and #includes it in topol.top
> apart from adding it into [ molecules ] section. So now its working but the
> question no. 3 is still on :)
>

I guess that works, but it's incredibly convoluted.  If you provide genion a
complete command line, then it should work fine, i.e.:

genion -s input.tpr -p topol.top -nn 5 -nname CL- -random

The above should be sufficient to insert 5 Cl- ions into your system, and update
your topology accordingly.

<snip>

> Again sorry for the confusion(and I know its not magic :)) but "apart" from
> providing the ion name to be added etc., according to the manual, you can
> add ions by following ways-
> 1) provide no. of ions to be added by -nn.
> 2) use -neutral so that if the system already has a net charge, that will be
> compensated.
> 3) use -conc.

To get (2) and (3) to work, simply specify -pname and -nname; if this is unclear
in the manual perhaps some additional information would be useful, but truly you
should probably always default to specifying these options yourself; the
defaults are for ffgmx, which is deprecated.  If you specify -pname and -nname,
then genion will always work properly for the above command flags.

<snip>

>
> > I can't comment on this specifically, but have you tried it?  Does it give
> > some
> > sort of error message if you do?
>
>
> I tried using above but grompp says its not possible. Meanwhile, I read in a
> paper(JCP2003(118),4692-4701) that they used a modified PME with 3 times
> bigger box dimension in the z-direction than the slab geometry. And I am
> using it with periodicity=xyz turned on.
>
> So unless I use walls, I can't use periodicity=xy with coulumb-type=PME and
> ewald_geom=3dc, Is that right?
>

Seems like that should be right; what is the exact error message that grompp is
giving you?  That is more informative than saying "its not possible."

-Justin

========================================

Justin A. Lemkul
Graduate Research Assistant
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul at vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin/

========================================