[AMBER] DFTB/MM result from SQM does not match Sander?

From: Ge Song via AMBER <amber.ambermd.org>
Date: Wed, 26 Mar 2025 14:30:27 +0000

Dear Amber developers,

I was trying to get the mulliken charges for the DFTB/MM calculation. In the beginning, I was using sander to do the DFTB/MM calculation, but it seems that sander does not support to print the mulliken charges information, so I turned to sqm. According to the manual, the code of sqm is extracted from sander, so I assume that the result of energy should also match with sander (as long as the settings are the same), with the ability to print the mulliken charges. However, I have tried to keep the settings same and do calculation for the same frames (capped alanine in solution with no periodicity), but the results do not match. I was wondering if this is normal, or something different need to be set for sqm.

My qmmm settings for sander look like:
 &cntrl
  imin=5,
  ntx=5,
  irest=1,
  cut=9999.0,
  ntb=0,
  ntwx=1,
  ntwf=1,
  ioutfm=0,
  ifqnt=1
 /
 &qmmm
  qmmask=':1-3',
  qm_theory='DFTB',
  dftb_disper=1,
  qmcharge=0,
  spin=1,
  qmshake=0,
  qm_ewald=0, qm_pme=0,
 /

I have extracted the coordinates, the MM charges (in a.u. or e), and my input for sqm looks like:
Run semi-empirical minimization
&qmmm
  qm_theory='DFTB',
  qmcharge=0,
  dftb_disper=1,
  maxcyc=0,
  qmmm_int=1,
 /
 1 H1 8.67362881 13.27917004 23.81789780
 6 CH3 9.10045052 13.75533867 22.93517876
 1 H2 8.48295784 14.52715492 22.47572708
... (QM coordinates)
#EXCHARGES
 8 O 1.53231990 10.79592514 15.66447353 -0.834000
 1 H1 1.96565151 10.58137131 14.83838463 0.417000
 1 H2 1.09105611 11.62814236 15.49438477 0.417000
... (MM coordinates and charges)
#END

The result from sander is: DFTBESCF= -2335.9716
The result from sqm is: Heat of formation = -2301.18474934 kcal/mol or Total SCF energy = -16443.26125335 kcal/mol

I have also tried to compare the relative energy between two frames, but the relative energy does not match neither... I have tried amber19, amber22, amber24, and they all give the same results.
But in the end, I would like to know if there could be a way to print the mulliken charges for DFTB/MM calculation that match with what have been done in sander?

Best,
Ge Song
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Mar 26 2025 - 08:00:02 PDT
Custom Search