AMBER Archive (2004)

Subject: Re: AMBER: Spike in energy and RMS

From: Andreas Svrcek-Seiler (svrci_at_tbi.univie.ac.at)
Date: Thu Sep 09 2004 - 04:37:57 CDT

Hi,
> I am minimizing a protein using AMBER 7 on IRIX platform. During
> minimization, the energy sharply rises at a single step and then returns
> to lower value on the very next step. Analysis shows bond and angle
> energy increases sharply as compared to other energies. And this is
> repeated at every 25th step. The output shows similar pattern of spike in
> energy and RMS with IGB=one as well. However, at the end, the minimization
> converges to a visually normal structure meeting convergence criteria.
>
> Is this behaviour normal? Can anyone explain it?
...Sometimes the minimization routine does a trial step, which
might be "too large", resulting in energy spikes. Whether (and how often)
that happens depends on the algorithm as well as on implementation
details, however it is normal and harmless.
regards
Andreas
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu