AMBER Archive (2009)

Subject: Re: [AMBER] Problems in the result of MM_PBSA

From: David A. Case (
Date: Fri May 08 2009 - 07:45:19 CDT

On Fri, May 08, 2009, Sean Zhao wrote:

> I used Amber 9 to simulate a small molecule bound in a protein. After I have
> finished the MM_PBSA, I found the result ( shown below ) was very
> unreasonable. For instance, the delta VDW, INT, GAS, PBTOT and GBTOT are all
> beyond 10 million. However, no error was reported in the terminal when I was
> running the MM_PBSA. How can I fix this issue? Any response will be
> appreciated.

> VDW -550.98 22.86 10037286.88 13997.12 840507.41
> INT 3434.99 36.17 10647284.69 171501.99 369629.85

Note that you have completely unreasonably van der Waals and internal
energies for the receptor. Something has gone very badly wrong with that
part of the simulation, and you need to find out what it is. Certainly,
visual examination is an absolute must here, as Ray pointed out. It has
nothing to do with some small difference between am1-bcc and RESP charges.

In general: the MM-PBSA scripts are intended to be a bookkeeping aid for
experienced users, who have well-behaved systems. Everyone should carry out
their first such simlations "by hand", doing each step themselves and making
sure they understand what is going on. Even for experienced users, the
scripts are weak in catching errors -- everyone who has problems needs to have
(or develop) the ability to critically examine simulations to look for


AMBER mailing list