[gmx-users] Membrane simulations

[Justin Lemkul]

On 10/20/14 11:40 PM, lloyd riggs wrote:
> Dear All,
> So I have been doing membrane simulations on a small (really small) node for a
> few months, and was able to get in a few hundred nanoseconds of time.  I now
> have added 4 inositol lipids, and repeatedly encounter some errors as follows.
> I set the time in the eq for 0.0005 and it runs about 25000 steps.  I then move
> up to a normal sim at 0.002 and it runs about 25000 steps then crashes with a
> particle moved to far error.  On inspection, these vary each time, but are
> always a membrane lipid /cholesterol or inositol atom, usually trying to move
> ithrough a periodic boundary temporarily?  I then run the sim again
> (continuation) for 25000 steps at 0.0005 again then it crashes again with the
> same eror.  I then Switch back to 0.002 as a continuation, and the same.  I have
> done this for around 10 ns now?  I woundered if there were any suggestions, as
> it makes it impossible to run a simulation unless you baby sit it constantly,
> and precludes submitting it to a server?  Please excusse my spelling.

It sounds like however you have added the inositol is not stable, otherwise the 
topology itself is not stable.  How did you add the inositol?  How well did it 
minimize?  What conditions are you using for equilibration (.mdp)?

-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

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