[AMBER] ntt=3 and ibelly=1

From: Kristina Furse <Kristina.Furse.1.nd.edu>
Date: Thu, 16 Jul 2009 17:07:23 +0100

Hi all-

I just did something stupid that cost me time and energy, so I thought I'd share it in the hopes it might prevent someone else from doing the same.

I've run a number of regular, explicit DNA simulations, and wanted to do some frozen DNA simulations, using ibelly=1, for comparison. I chose a set of snapshots, centered the DNA, imaged the water and ions around it, then did a quick minimization (250 steps) to clean up any instantaneous weirdness in the DNA structure since I was going to freeze it. I run my simulations NVE, but did some NVT equilibration to re-establish the temperature for the water and ions with the DNA frozen. That was my problem--in my general equilibration protocol, I use Langevin dynamics, ntt=3. Apparently this is not compatible w/belly. My DNA atoms still got kicked around during the equilibration even though the moveable belly only included water and ions. I ended up with hydrogens frozen in very strange positions for my subsequent NVE trajectories. Temps and energies looked fine--bond and angle energy didn't show up in the output b/c of belly--so I only caught it by visualizing the trajectory to hunt down a different problem.

So, long story short, no ntt=3 with ibelly=1, unless I'm missing some other flag or setting that would have rescued the plan.


Kristina Furse
Postdoctoral Research Associate
260B Stepan Chemistry Hall
Notre Dame, IN 46556
AMBER mailing list
Received on Thu Jul 16 2009 - 10:09:48 PDT
Custom Search