Try a restart.
If you do a restart on a const-P simulation with restraints, the
reference structure gets set back to the UNSCALED coordinates. Thus
the restart doesn't exactly pick up where the last simulation left
off. Instead, the (possibly scaled) simulated structure suddenly
finds itself being pulled toward the unscaled reference structure.
For some purposes, this is a bug but for yours and many others, I
think it's a feature. Your restarted simulation will suffer some
weird T-flucturation, but then it will get pulled back towards
structural sanity, assuming the second simulation doesn't do too much
drastic rescaling, in which case, restart again.
Of course, the other thing that SHOULD have kept your structure fairly
sane in spite of refcoord rescaling are things like bond stretch and
angle forces. But your restraint forces are so strong that they
might overpower those. I'd suggest that if const-P is likely to
result in big rescaling, be more gentle with the restraints during
pressure equillibration.
-Don
At Fri, 10 Sep 2010 12:26:41 -0500,
Sergio R Aragon wrote:
>
> Hello Alexander,
>
> I normally use a restraint parameter of 500 when the intent is to keep the protein from moving. What parameter do you use? Can you increase it? The other thing to make sure of is that the number of residues to hold fixed does include the ligand portion of your molecule. Could it be that you are not fixing the atoms of the ligand and only the protein?
>
> Cheers, Sergio
>
>
> -----Original Message-----
> From: Alexander Metz [mailto:alexander_metz2000.yahoo.de]
> Sent: Friday, September 10, 2010 3:41 AM
> To: amber.ambermd.org
> Subject: [AMBER] Complex distorts during pressure adjustment in equilibration
>
> Hello,
>
> I am currently equilibrating a system for MD like many times before. The problem
> is that during the NTP pressure adjustment step my complex gets distorted. This
> is due to the distance rescaling. It has not been a big deal for globular
> proteins with the ligand rather close to the center of the solvent box since the
> rearrangement of the complex was minimal.
>
> But the system I want to study is rather elongated. I use a rectangular box
> (~130x40x40A). Also the initial density is rather low after solvating with LEaP
> (~0.8g/ml). The ligand resides at one end of the elongated protein.
>
> This results in the ligand shifting significantly (>2A) with respect to the
> protein during the density adjustment. Also the 2 parallel helices in the
> dimeric protein shifts significantly. This happens depsite of using positional
> restraints.
>
> In general I think that only the water should adjust so that that solute stays
> in its original configuration. But unfortunately the positions of all molecular
> entities are rescaled.
>
> Is there an easy way to prevent this?
>
> Best regards,
>
> Alexander
>
> ++++++++++++++++++++++++++++++++++++++++++++++++++
> Alexander Metz (PhD Student)
> Heinrich Heine UniversitätDüsseldorf
> InstitutfürPharmazeutischeundMedizinischeChemie
> Computational Pharmaceutical Chemistry
> Universitaetsstrasse 1
> 40225 Düsseldorf, Germany
> Tel: +49 211 81 13854
> ++++++++++++++++++++++++++++++++++++++++++++++++++
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
Email Disclaimer: www.stjude.org/emaildisclaimer
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Sep 10 2010 - 17:00:03 PDT
