Re: [AMBER] Nudged Elastic Band solvent choice and initial heating problem

From: David A Case <>
Date: Wed, 3 Feb 2021 09:00:45 -0500

On Wed, Feb 03, 2021, 李明玉 wrote:
>I am trying to perform nudged elastic band method (NEB) on a large
>1618-residue protein.
>I'm new to MD.

I don't use NEB, so can only provide some general advice here: start with a
*much* smaller system as a test case, and see if you can get explicit
solvent to work. You'll have a much better chance of finding problems, and
if you still have problems, a system that others might be willing to look at
to reproduce the behavior that you see.

If you find that you need a time step of 0.0005, something may be
amiss with your system setup. Try a short simulation (say nstlim=100 and
ntpr=1) and see if that helps. Are you sure you system is well-minimized?
You can use the "checkstructure" action in cpptraj to look for common

If you get the small system to work, you can scale up. But you might still
think if you need just a big system. Most proteins that have 1600 residues
are made up of smaller domains? Might you not learn what you need to know
by looking at some sub-system?


AMBER mailing list
Received on Wed Feb 03 2021 - 06:30:03 PST
Custom Search