[gmx-users] First frame already out of box, getting very large RMSD

[Justin Lemkul]

On 2/28/15 7:38 PM, yunshi11 . wrote:
> On Sat, Feb 28, 2015 at 4:21 PM, Justin Lemkul <jalemkul at vt.edu> wrote:
>
>>
>>
>> On 2/28/15 7:17 PM, yunshi11 . wrote:
>>
>>> On Sat, Feb 28, 2015 at 3:03 PM, Justin Lemkul <jalemkul at vt.edu> wrote:
>>>
>>>
>>>>
>>>> On 2/28/15 6:00 PM, yunshi11 . wrote:
>>>>
>>>>   Dear all,
>>>>>
>>>>> I am running MD for a protein-ligand complex in a dodecahedron box and
>>>>> followed the "Suggested trjconv workflow" from
>>>>> http://www.gromacs.org/Documentation/Terminology/
>>>>> Periodic_Boundary_Conditions
>>>>> .
>>>>>
>>>>> Now I wonder how to remove jumps (across periodic boxes) when the first
>>>>> frame (actually the .gro file from equilibration run) is already jumping
>>>>> across two (or more) boxes (according to visualization in VMD).
>>>>>
>>>>> I understand that this is just an visualizing artifact, but it seems to
>>>>> also affect other analyses such as calculations of RMSD along the
>>>>> trajectory. I got some impossible RMSD values like 1.5 nm, considering
>>>>> other replicate simulations give only ~0.3 nm.
>>>>>
>>>>> Any idea on how to fix this (the effect on RMSD calculations etc., not
>>>>> visualization)?
>>>>>
>>>>>
>>>>>   trjconv -center -pbc mol -ur compact usually solves all problems for
>>>> such
>>>> simple systems.  Center on the protein, everything else gets re-wrapped
>>>> around it. More complex operations would be needed for protein complexes,
>>>> membranes, etc. But proteins in water (with or without ligands) are
>>>> generally straightforward.
>>>>
>>>>
>>>>   I did try that, and other combinations of -pbc and -ur options. Sorry
>>> that
>>> I should have been more specific:
>>>
>>> My system is a homo dimer of protein, each with two ligands (cofactor +
>>> inhibitor) bound symmetrically. Some trjconv options can put the dimer
>>> protein together, but the two ligands of the second monomer are always in
>>> the other periodic boxes.
>>>
>>>
>> This changes the approach considerably; always provide full details in the
>> first message - you'll arrive at a solution faster!
>>
>> The centering in this case should be done with respect to one protein or
>> some chosen custom group of residues that makes a logical center (e.g. the
>> interfacial residues of one subunit).  The same concept applies - the
>> remaining elements of the system (the other protein, the ligands, etc) will
>> be wrapped with respect to the centered subunit.
>>
>>
> With trjconv -s x.tpr -f x.gro -n x.ndx -center -pbc mol -ur compact -o
> x.gro, I tried centering on monomer A, monomer B, and two interfacial
> residues of monomer A. None of them worked, and the two ligands of monomer
> B are always outside the box.
>
> Can I send .gro files to the mailing list?
>

No. To share files you need to upload them to a file-sharing service and provide 
a link.  The input files (.tpr, .gro, and .ndx) would be useful.

-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

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