I think that you are right about the problem.
The best-fit and RMSD calculations do not
image the system before comparison to the
reference coordinates. This would certainly
cause large energies and forces.
I am not sure why this happened on restart and
not during the run, it should have happened right
when sander imaged one of the strands. I
will have to look into this and see what is happening.
For now it is probably best not to use imaging
when you are using tgtmd=1.
Carlos
----- Original Message -----
From: <raist.mmb.pcb.ub.es>
To: <amber.scripps.edu>
Sent: Monday, November 17, 2003 1:54 PM
Subject: AMBER: Restarting MD amber7
> Hi all,
> the other day I posted a mail regarding problems restarting MD with
> amber7. I was running TGMD on a DNA+ions+water and after 8ns sander
> complained with the usual SHAKE problem (when visualizing the structure it
> was fine but the 2 strands were not imaged together). I checked the
> energies to see there were no flying ice cube or other artifacts during
> the simulation.
>
> When I image the last step of the previous run and start the simulation
> from there (with the corresponding velocities) sander does not complain.
> It does not complain either when I use the non-imaged restart file turning
> off TGMD and using amber 7 and amber 6. So I guess the problem comes from
> TGMD when you implement the restrains to target atoms and it needs to
> compare to the reference structure. Has anyone had similar problems??? Do
> you think this is really the problem??
>
> Thanks in advance.
>
> ----------------------------------------------------------------
> Alberto Perez
> Molecular Modeling & Bioinformatics group
>
> IRBB Parc Cientific de Barcelona | e-mail:raist.mmb.pcb.ub.es
> Universitat de Barcelona | phone: + 34 93 403 71 55
> Josep Samitier,1-5 | FAX: + 34 93 403 71 57
> 08028 Barcelona (Spain) |
>
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Mon Nov 17 2003 - 12:53:01 PST
