AMBER Archive (2006)

Subject: Re: AMBER: Frozen at Minimization

From: Chris Moth (chris.moth_at_vanderbilt.edu)
Date: Tue Dec 19 2006 - 09:01:16 CST

I routinely minimize a solvated 1000 residue protein on my desktop system.
Each mininimization step does not take that long.

If I were faced with your results I might try to home in on the trouble
with some rather unintellectual diagnostic steps:

1) Restraining a smaller number of residues, for grins. Does the problem
appear as soon as you restrain one or two residues?

2) Triple check restraint masks to make sure you are not accidentally
restraining side chains or waters.

3) Request output at EVERY step (ntpr=1 perhaps). What appears in the
minout file?

4) I might try a smaller restraint energy - say 10kCal - just to see what
happens

5) If I had access to resources to run PMEMD, I might try the minimization
with it.

More advice might depend on getting more details of your system and input
parameters. I don't think this will turn out to be a "major problem" for
you.

C

> Hello
>
> I am attempting an energy minimization on a large protein ( greater than
> 1000 residues). When I do energy minimization without any restraints
> AMBER does minimize the structure although it takes a while (running
> maxcycle=200). However, when I put in restraints on the backbone (100
> kcal/mol) something weird starts to happen. AMBER, with restraints
> applied, runs but seems to "freeze" up in the sense that I have been
> running minimization for over 24 hours and the output file still reads
> the same as it read just a few minutes into the calculation. It freezes
> at the point where it lists the group and number of atoms for restraint.
> Does this mean that AMBER is frozen at a step in the calculations? I am
> afraid that this is a response in AMBER due to very high potential
> energy values. Is this correct?
>
>
>
> Thanks, Steve
>
>
>
>

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu