[gmx-users] the protein drifts when molecular dynanic simulation

[Choon Peng]

Dear Florian and zjim,

   The problem that zjim faces could be due to overall translation of the
protein. 
I don't know how that can be fixed during post-processing but the way to
remove overall translation during MD is to have:
comm_grps = Protein
comm_mode = Linear

in your mdp file.


   As for artifacts caused by PBC, after much 'pain-staking' experimentation
I found that the following works with trjconv: (for v3.2.1 and above)

-pbc whole -ur compact -center

This makes the protein whole and puts it in the center of the octahedron or
whatever shaped solvation box you specified.
Especially important for us as we're interested in buried water molecules.

However, rotations of the protein is still possible.
To fix rotations:
-fit rot+trans

But you can't use this with "-pbc whole -ur compact -center" at the same
time.


A way to resolve both problems is to define in your md.mdp:
comm_grps = Protein
comm_mode = Angular

This will remove overall rotation and translation of a protein.
You'll get a warning that this works only for a single molecule.


Regards,
Choon-Peng
-- 
Mr Choon-Peng CHNG
Research Associate
Computational Biology Group
Bioinformatics Institute, A*STAR
30 Biopolis Street, Level 7. Singapore 138671.
DID: +65 6478 8301 FAX: +65 6478 9047
WWW: http://www.bii.a-star.edu.sg/~cpchng


On 11/30/04 9:15 PM, "Florian Haberl"
<Florian.Haberl at chemie.uni-erlangen.de> wrote:

> Hi
> 
> On Tuesday 30 November 2004 13:39, zjim wrote:
>> Dear gmx-user
>> When perform protein md in a box with, the protein drifts to one side of
>> the box,and as the md ends(6 ns),some of it even comes out the box,this
>> maybe a problem. so my quesition is why is the protein drift when
>> simulation and how to avoid this. Somebody maybe help me
>> thank you in advance
> 
> perhaps you got same prob i got:
> 
> http://www.gromacs.org/pipermail/gmx-users/2004-November/012902.html
> 
> should help you
> 
> Greetings,
> 
> Florian
>