Re: [AMBER] Analysis of minimization stage

From: Jason Swails <jason.swails.gmail.com>
Date: Thu, 15 May 2014 11:46:54 -0400

On Thu, 2014-05-15 at 09:35 -0600, Daniel Roe wrote:
> On Thu, May 15, 2014 at 7:07 AM, Valentina Romano
> <valentina.romano.unibas.ch> wrote:
> > BOND = 7906.5852 ANGLE = 529.6831 DIHED = 2310.5740
> > MD
> >
> > BOND = 120.7537 ANGLE = 529.6831 DIHED = 2310.5740
> >
>
> This appears to be your problem: your bond energy jumps from 7906.5852
> to 120.7537. All of your other energy components are the same. Are you
> sure you are using the same topology file?

This looks to me like a SHAKE problem. You are _not_ using SHAKE for
minimization (ntc=1, ntf=1 by default), but then you set ntf=2 and
ntc=2, which disables the energy calculation of the SHAKEn bonds.

The very high bond energy from the minimization suggests (to me!) that
some of the bonds that are constrained in the MD adopt very strained
configurations during the minimization (for whatever reason). This
would explain why you get SHAKE failures when you start dynamics.

You can try setting ntc=2 and ntf=2 for the minimization as well and see
if that helps fix things. Note that SHAKE is not strictly correct for
minimization, but if you are just looking to relax your structure to
start dynamics it is typically OK.

Good luck,
Jason

-- 
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu May 15 2014 - 09:00:10 PDT
Custom Search