[gmx-users] Umbrella Sampling gromacs 5.0 error

[Justin Lemkul]

On 11/23/14 3:56 PM, Alexander Law wrote:
> Hi Justin
>
> I downloaded the compressed folder from this website: http://www.gromacs.org/Documentation/Tutorials/GROMACS_USA_Workshop_and_Conference_2013/Computing_potentials_of_mean_force%3A_Justin_Lemkul,_Session_2A
>
> Within it the the npt_umbrella.mdp file reads continuation = yes, gen-vel = yes.
>
> I've ran ~15 simulations over the weekend with continuation = yes, gen-vel= no
>
> Are these simulations unsuitable for umbrella sampling?
>

The settings should be somewhat intuitive; if you are continuing a run (i.e. 
some previous time interval or from equilibration), you should not generate 
velocities (continuation = yes, gen_vel = no).  If you are starting a new 
simulation, i.e. equilibration (continuation = no) then you should generate 
velocities (gen_vel = yes).

I have checked twice since you originally posted this problem, and 
npt_umbrella.mdp provided with the tutorial (for equilibration) is correct 
(continuation = no, gen_vel = yes) and the input for umbrella sampling 
(md_umbrella.mdp) is also correct (continuation = yes, gen_vel = no), so I don't 
know what this issue related to the contents of npt_umbrella.mdp is.

> ---------------------------
> I am assuming they are, so I've started a new sequence with not_umbrella.mdp reading the same as: http://bevanlab.biochem.vt.edu/Pages/Personal/justin/gmx-tutorials/umbrella/Files/npt_umbrella.mdp
>
> This produces this error:
>   WARNING 1 [file npt_umbrella.mdp]:
>    You are generating velocities so I am assuming you are equilibrating a
>    system. You are using Parrinello-Rahman pressure coupling, but this can
>    be unstable for equilibration. If your system crashes, try equilibrating
>    first with Berendsen pressure coupling. If you are not equilibrating the
>    system, you can probably ignore this warning.
>
> Since I am not equilibrating I've avoided this using the -maxwarn command and started a new sequence for umbrella sampling.
>

Again, npt_umbrella.mdp *is* intended for equilibration; the warning probably 
shouldn't be a fatal one for grompp, but it is indicating that instability may 
result.  If the simulations proceed, that's fine.  The warning should probably 
just be a note, instead, but in general one should not generate velocities with 
Nose-Hoover, which allows for wider fluctuations than a weak coupling method 
like Berendsen or V-rescale, which are more robust for equilibration.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

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