Re: Pointcharge file in QM/MM setup with ORCA

From: Acharya, Atanu (aacharya42_at_gatech.edu)
Date: Mon Apr 08 2019 - 09:18:58 CDT

Dear Marcelo,
   Thank you very much for your suggestions. Indeed, the equilibration and subsequent MD steps look good and stable.
For some reason, if the minimization performed for a large number of steps (~1000) the QM subsystem goes to some distorted geometry.
Running the minimization for ~200 steps solved the problem.

Best,
Atanu

________________________________
From: Marcelo C. R. Melo <melomcr_at_gmail.com>
Sent: Thursday, April 4, 2019 4:05 PM
To: NAMD; Acharya, Atanu
Subject: Re: namd-l: Pointcharge file in QM/MM setup with ORCA

Hi Atanu,
Yes, the point charges are (by default) re-defined at every step. Using the cutoff distance, NAMD selects a "shell" of point charges around your QM region(s).
You can provide a PDB file with a static selection of point charges (check the qmCustomPCSelection keyword in the documentation) if maintaining a constant selection is important for your particular case.
As for minimization, usually the QM energy is much greater than the MM energy, and tends to dominate the potential. Not knowing more about your case I could only suggest running an equillibration to see what happens.
Best,
Marcelo

---
Marcelo Cardoso dos Reis Melo, PhD
Postdoctoral Research Associate
Luthey-Schulten Group
University of Illinois at Urbana-Champaign
crdsdsr2_at_illinois.edu<mailto:crdsdsr2_at_illinois.edu>
+1 (217) 244-5983
On Wed, 3 Apr 2019 at 15:06, Acharya, Atanu <aacharya42_at_gatech.edu<mailto:aacharya42_at_gatech.edu>> wrote:
Hi there,
 I am trying out the NAMD QM/MM interface with ORCA. I noticed that the  .pntchrg file is written in the qmBaseDir directory.
However, the number of pointcharges in the file is different in each step. Is this because NAMD is writing the charges within the cutoff distance?
In my calculations, I noticed the that during QM/MM minimization the QM energy does not go down (the MM energy does go down though).
Can someone provide some insight into that?
Thanks a lot,
Atanu

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2020 - 23:17:10 CST