Dear AMBER Community,
For the QM/MM simulation of a covalently bound ligand to a Ser
residue, we considered the Ser-ligand as a modified amino acid (SEN).
Following Antechamber, the atom type for a sulfur present in the ligand
part was given as "SO".
ATOM 20 SAR SEN 46 -15.313 1.434 -29.065 2.849000 SO
While trying to run a QM/MM simulation, we got the following error in the
output file for minimization-
3.1 QM CALCULATION INFO
--------------------------------------------------------------------------------
QMMM: Atom number: 26 has atomic number -1.
QMMM: There are no PM3 parameters for this element. Sorry.
SANDER BOMB in subroutine qm2_load_params_and_allocate
UNSUPPORTED ELEMENT
QM PM3 NOT AVAILABLE FOR THIS ATOM
This error pertains to the sulfur atom, and its coordinates in the
minimization output file were written as following -
QMMM: QM_NO. MM_NO.* ATOM* X Y Z
QMMM: 26 710 *^.^.* -1.0743 -0.8644 -2.3790
The input file for minimization contained the following, which we obtained
from the AMBER tutorial for QM/MM MD (
https://ambermd.org/tutorials/advanced/tutorial2/section3.php) -
Initial min of our structure QMMM
&cntrl
imin=1, maxcyc=500, ncyc=200,
cut=8.0, ntb=1, ntc=2, ntf=2,
ifqnt=1
/
&qmmm
qmmask=':46',
qmcharge=-1,
qm_theory='PM3',
qmshake=1,
qm_ewald=1, qm_pme=1
/
I tried with some other qm_theory such as 'PM6' and 'AM1', which gave the
same error. *A conventional all-atom MD*, however, *ran successfully with
the same prepared files*. Is the error because of the "SO" atom type? I
would be grateful for any guidance on this topic. On another note, is
running a QM/MM simulation on a GPU using *pmemd.CUDA *possible? Is there
any way for faster calculation than sander?
Thank You
Sincerely,
Sunanda Samanta
MSCPh.D.
BSBE, IIT Indore
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Aug 04 2024 - 23:30:02 PDT
