[gmx-users] POPC simulation with heavy hydrogens and 4 fs timestep

[Justin Lemkul]

On 11/19/15 7:49 AM, Ganesh Shahane wrote:
> Dear Gromacs Users,
>
> I have been trying to simulate a pure POPC bilayer with a timestep of 4 fs
> and heavy hydrogens, using the charmm36 ff. Towards this end, I constructed
> the bilayer using Charmm-GUI and used pdb2gmx with the -heavyh flag to
> generate the appropriate topology.
>
> However when I use the grompp to generate the .tpr file, I get a warning:
> "Overriding atomtype OT". A quick look in the Charmm36 ffnonbonded.itp
> revealed that the atomtype OT corresponds to an oxygen with reduced atomic
> mass, which is perfectly understandable given the use of -heavyh flag.
>
> My problem is that I don't understand what the warning exactly means. A
> look at a previous gromacs thread where this "overriding atomtype" problem
> has been reported, didn't help.
>
> Is it safe to ignore this warning?
>

In this case, yes, because the atom type overriding the previous instance has 
the same LJ parameters, just a different mass.

> When I ignored the warning and carried out a simple NVE run, it gave me
> potential, kinetic and total energies that were not conserved. Please find
> the plot of the energies and the associated NVE run parameters at this link
> https://drive.google.com/file/d/0B2cmuCsErAPZcjhmNW1tWnA3UXM/view?usp=sharing.
> A look at the trajectory in VMD revealed that the molecules which were
> initially very dynamic, start "freezing" as the KE approaches zero.
>
> Does anybody has any idea as to where I am going wrong?
>

I would not expect any simulation with a 4-fs time step, mixed precision, 
constraints, etc. to conserve energy well under NVE.  But the time you're 
showing here is totally inadequate to evaluate conservation of energy, anyway. 
Normally one would equilibrate for some time under NVT, then switch to NVE. 
This issue gets discussed all the time on the list, so check the archive for 
recent commentary.

-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

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