[gmx-users] analysis of different groups with g_energy

[Peyman Yamin]

On Tuesday 10 June 2008 16:07, Mark Abraham wrote:
> Peyman Yamin wrote:
> > Good day!
> >
> > Thanks for the comments, but maybe it could be done in a less laborious
> > way as following?
> >
> > g_energy gives all the interactions between MOL and SOL other than
> > bonding ones. since there is no bond between MOL and SOL, one could do a
> > ONE step simulation of MOL with its last acquired structure from solvent
> > in vacuum and - since it's the whole system now - use g_energy again to
> > get Angle and Dihedral energies. The only problem would then be the Bond
> > itself (stretching) which is not important since when the MOL itself is
> > the whole system, one could even use the tot-energy directly from the
> > one-step MD in vac and by substracting get even the stretching energy.
>
> That seems at least roughly equivalent. Laborious is a subjective
> judgement and depends what technique you would use to do it :-)

The world is subjective, to a great extent! ;)

>
> > should this be at least OK in a qualitative way. (?)
> >
> > Despite of a good discussion on the topic (march,2008 - David, Xavier,
> > Pascal, Berk), I still can't find what contributes exactly to SR,LR and
> > 1-4 non-bonding interactions and what does not, other that what they
> > stand for ;)
>
> I said last time that the place to look for how these might be working
> was in the early chapters of the GROMACS manual. The .mdp options you're
> actually using will vary the manner in which these are calculated.
>

I know I know, but honestly there is not that much detail there! I think I 
should look at the code! and it's what I mostly try to avoid! But gmx is so 
nicely written that I've not done that since a long time - and become lazy!

> > are they really independent and additive to give the whole interactions?
>
> Did you try adding them up?
>

Well yeah, and I get reasonable results! Qualitatively great! correlations are 
good, I have trends for interacions which are very good corresponding the 
other - not fully dependent - dynamic parameters, like those having a size 
feel - prediction of entropical trends for example, and their comparison with 
some simple theories gives beautiful results which make me xmgracing since 
some hours!  



> > or
> > for example the Reaction Field corrections appliy to some and to some
> > not? But as far as I FEEL, they are in a good way, so that one can do as
> > above (at least with a reasonably small error).
> >
> >
> > If it's not supposed to work from the beginning, i.e. the terms are not
> > additive at all, please let me know! as well as if one should
> > add/substract some things!
>
> First, try to understand the algorithm you're implementing. Then you
> will have a much better idea what quantities are what.
>

It's true. I try to do this way mostly!

Peyman

> Mark
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php

-- 
Peyman Yamin
Lehrstuhl fuer Thermische Verfahrenstechnik
Universitaet Erlangen-Nuernberg
Egerlandstr. 3
91058 Erlangen
Phone: +49(0) - 9131 - 85 27671
Mailto: peyman.yamin at cbi.uni-erlangen.de