[gmx-users] spliting trajectory g_hbond -contact

[Justin Lemkul]

On 2/8/14, 5:32 AM, Kavyashree M wrote:
> Dear users,
>
> I need a small clarification . I have to get an index file for the
> contact between protein and water with a cut of distance of
> 0.35 nm. its a 20ns run. When i use g_hbond along with -contact
> it runs for a while, say few ns and gets killed due to the memory
> overload. The system I use has 16Gb RAM and i cannot increase it.
>
> So the question is instead of calculating the contact between the
> two entities for the whole trajectory, if I split the trajectory into 2
> halves and then calculate the contact, will the combined index
> file obtained be same as the one had I used the whole trajectory?
>

My gut reaction is to say no.  The index file tracks the interactions that exist 
in the order that they show up, hence they can be mapped to the .xpm file 
produced by -hbm.  The interactions at t=0 and t=sometime later will not be the 
same, so the index file will have a different order.  Consider reducing the 
number of trajectory frames, or ordering the waters with trjorder and only save 
the N closest waters that actually matter.

-Justin

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

Justin A. Lemkul, Ph.D.
Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

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

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