Re: [AMBER] QM/MM using Amber and Gaussian: Discrepancies between the forces from the gaussian output file and the forces reported in the mdfrc file

From: Goetz, Andreas via AMBER <amber.ambermd.org>
Date: Mon, 11 Nov 2024 02:34:42 +0000

Dear Ge Song,

If this is from a QM/MM calculation, then the difference in forces is due to Lennar-Jones interactions.

If you add “debug = 2,” to your &gau namelist, then the QM/MM interface will print the converted energy in kcal/mol and forces in kcal/mol/Angstrom to mdout. This should match with the numbers that you converted manually as it does not contain LJ energy or forces. In a pure QM calculation (no MM atoms) the forces should match with the content of the mdfrc file.

All the best,
Andy


Dr. Andreas W. Goetz
Associate Research Scientist
San Diego Supercomputer Center
Tel: +1-858-822-4771
Email: agoetz.sdsc.edu
Web: www.awgoetz.de

On Nov 10, 2024, at 5:57 PM, Ge Song via AMBER <amber.ambermd.org> wrote:

Dear Amber community,

I am running single point QM/MM in Sander with forces being output. The calculation finished normally, and the energy from the gaussian log file and that from the mdout file are consistent according to https://urldefense.com/v3/__http://archive.ambermd.org/201812/0031.html__;!!Mih3wA!ASmoTRVtoWAjX-zZbgwTGW2sjRB29QG5fnInvrtGLFnlteLbQs_Wx-vZ6A4_VqarXn9HFu3i-kN-NA$ . However, when I also checked the forces, I found they are roughly consistent but significant discrepancies exist:

The forces on the first atom in the gaussian log file (old.gau_job.log) is:
      -0.012109279 -0.000859028 -0.003084640
those are in Hartree/Bohr, converting into kcal/mol would be:
      -14.359     -1.019      -3.658

While in my mdfrc file, they are:
      -14.517     -0.793      -3.978

So, they look similar, but I don't think these errors are neglectable.

I have tried both Amber19 and Amber22, and they gave the same results. Here is my input setting of sander:

Production
&cntrl
 imin=5,
 ntx=5,
 irest=1,
 temp0=300.0,
 cut=9999.0,
 ntb=0,
 ntwx=1,
 ntwf=1,
 ioutfm=0,
 ifqnt=1
/
&qmmm
 qmmask=':1-3',
 qm_theory='EXTERN',
 qmcharge=0,
 spin=1,
 qmshake=0,
 qm_ewald=0, qm_pme=0,
/
&gau
num_threads = 16,
mem='16GB',
use_template=1,
/

And the command line I was using:
mpirun -np 4 sander.MPI -O -i qmmm_rerun.in -o qmmm_rerun.out -p ala_amber.top -c amber_heat.ncrst -y amber_md_frame1.nc -frc mdfrc

Ge Song
Personal Academic Website of Ge - Physics of Life<https://urldefense.com/v3/__https://calvinge.github.io/gesong.github.io/__;!!Mih3wA!ASmoTRVtoWAjX-zZbgwTGW2sjRB29QG5fnInvrtGLFnlteLbQs_Wx-vZ6A4_VqarXn9HFu2ohTHtjQ$ >


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!Mih3wA!ASmoTRVtoWAjX-zZbgwTGW2sjRB29QG5fnInvrtGLFnlteLbQs_Wx-vZ6A4_VqarXn9HFu1cm7SqMA$

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Nov 10 2024 - 19:00:02 PST
Custom Search