Re: AMBER: total energy leaps

From: Carlos Simmerling <>
Date: Wed, 08 Jun 2005 07:22:06 -0400

with ntr and restarts you need to be careful what you use
as the reference coordinate. Do you keep the same -ref each time,
or do you change it? If it changes, then the restraint energy for the first
step will not match that for the last step of the previous run.
Examining the energy components (bond, angle, etc) may help
you see where the discontinuity comes from. It should work fine.
cristian obiol wrote:

>Dear Amber users,
>I am carrying out a NVT dynamics with shake=2 and restraints (ntr=1), I send
>10 short steps of 100ps. The problem occurs when I represent energy vs time,
>in a few first picoseconds the total energy does not continue with the
>previous value normally but gives a big leap, each 100ps.
>I solved this problem simply sending my total 1ns simulation in only one
>step, but this generates large outputs.
>I noticed that this problem does not appear when the dynamics not presents
>shake or restraints, it seems that the two variables (ntc and ntr) are not
>fully compatible ? How can I solve these first energy leaps in each 100ps
>step ?
>Thanks in advance,
>Cristian Obiol
>Biological Systems Modelling
>and Drug Design Research Group
>University of Barcelona
>The AMBER Mail Reflector
>To post, send mail to
>To unsubscribe, send "unsubscribe amber" to

Carlos L. Simmerling, Ph.D.
Associate Professor           Phone: (631) 632-1336
Center for Structural Biology Fax:   (631) 632-1555
Stony Brook University        Web:
Stony Brook, NY 11794-5115    E-mail:
The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to
Received on Wed Jun 08 2005 - 12:53:00 PDT
Custom Search