[gmx-users] Problem with trjconv to keep molecule whole for multi-chain protein

[Justin Lemkul]

On 3/27/14, 12:16 PM, arrow50311 wrote:
> Hi,
>
> I recently got one problem with trjconv in order to keep protein whole
> during post-simulation processing. The problem is my protein is a
> multi-chain protein. I used dodecahedron for explicit solvent simulation.
> When I use trjconv to process the data, the flag -pbc nojump -ur compact
> will still leave part of the protein in its periodic image, and -pbc
> molecule -ur compact will make each chain of the protein whole and in the
> visualized box but sometimes cannot keep all chains together.
>
> Could any experienced people know how to deal with that? Is there any way to
> keep all chains together as in the native protein just to ensure a
> continuous visualizing trajectory?
>

One pass with trjconv is almost never enough, especially with complex systems. 
In general, you'll likely need several rounds of periodicity treatment (remove 
jumps, then make molecules whole, etc) then maybe translational fitting and/or 
centering.  Often one needs custom index groups that specify residues at some 
convenient location (protein-protein interfaces, perhaps) to center on those 
rather than centering on some generic default group.

http://www.gromacs.org/Documentation/Terminology/Periodic_Boundary_Conditions#Suggested_trjconv_workflow

-Justin

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

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

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