On Tue, Nov 15, 2022, Nitin Kulhar via AMBER wrote:
>
>I performed a solvent minimization on a protein system....
First, try using a CPU if you have not done so already.
Look at the Amber energy as a function of step number (you may have to
decrease ntpr). If it looks like it is stabilizing, try a short MD run,
with ntt=3, gamma_ln=10., dt=0.002 (or even 0.001). See if you can get
things going that way.
The underlying problem: Minimization doesn't deal properly with SHAKE.
We have a fix for the next release (still being tested, I hope, by Amber
developers), but for now, you may need to restrict minimization to just
removing really bad forces.
You can try some minimization without SHAKE -- that has its own gotcha's,
but sometimes helps. Bottom line: you want to get to MD as soon as you can.
Don't be afraid to try some now.
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Nov 15 2022 - 08:00:04 PST
