[gmx-users] Re: gmx-users Digest, Vol 59, Issue 138

[Pawan Kumar]

Hello Justin Sir,

Greetings from Pawan
Thanks for your valuable suggestion and reply.
Initially I gave the emtol of 1000 and the output I got was :
Steepest Descents converged to machine precision in 163 steps but did not
reach the requested Fmax<1000.
Potential Energy = - 4.4516497e+05

Even I tried to minimize the popc bilayer which I took from Tieleman sir's
website ( before generating a bigger bilayer using genconf ) but that also
converged to machine precision but not to the requested Fmax<1000. How do I
proceed further ?
Thanks for your suggestions and help.

Thanking you,
Pawan

On Thu, Mar 19, 2009 at 9:13 PM,

> Pawan Kumar wrote:
> > Hello Justin Sir,
> >
> > Greetings from Pawan
> > Thanks for your valuable suggestion and reply.
> > After inserting the protein in the bilayer using genbox I have minimized
> > the whole system without using any position restraints (i.e. define =
> > -DFLEXIBLE in em.mdp file). I used vanderwaal's distance parameter (
> > -vdwd of 0.6 ) in the genbox step.
> > After running mdrun for energy minimization I got the output as :
> > Steepest Descents converged to Fmax<2250 in 14 steps.
> > Potential energy = - 6.9484700e+05
> > Maximum force = 2.2114819e+03 on atom 34277
> > Norm. of force = 5.0103039e+04
> >
>
> You should try for an Fmax of no greater than 1000.  2250 is still very
> high.
>
> > I tried decreasing the emtol value in the em.mdp file but it ended with
> > machine precision.
>
> How far did it converge?  What was Fmax?
>
> > I have read in literature that 5000 steps of Steetest Descents run is
> > required after inserting the protein in the bilayer which should be
> > followed by atleast 1000 steps of conjugate gradients. How can I
> > accomplish this ? Is there any parameter to be given in the em.mdp file
> > ? I use steep as the integrator in the mdp file for energy minimization.
>
> Read the manual.
>
> Whether or not that exact setup is going to be "required" is likely
> system-specific.  I would say that as long as your system converges to a
> stable,
> negative Epot with a reasonable Fmax (less than 1000, but ideally lower)
> then
> you *may* have an appropriate starting structure.
>
> -Justin
>
> > Please help with some suggestions.
> > Thanks in advance.
> >
> > Thanking you,
> > Pawan
> >
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20090320/1e5d7217/attachment.html>