On Thu, Apr 11, 2013 at 8:15 AM, dbaogen <dbaogen.gmail.com> wrote:
> Dear all,
>
> In our system, the Magnesium ions are included, so I try to
> use QM/MM-GBSA method to obtain much more accurate result. The receptor is
> included one Magnesium ion and protein, and the ligand is ATP with another
> Magnesium ion. The QM/MM-GBSA parameters are set as follows:
> &general
> startframe=1, endframe=100, interval=5,
> keep_files=0,verbose=1,search_path=1,
> receptor_mask=":1-916:918", ligand_mask=":917:919"
> /
> #&pb
> #istrng=0.1,
> #
> &gb
> saltcon=0.1,ifqnt=1,qmcharge_com=0,qmcharge_rec=2,
> qmcharge_lig=-2,qm_theory='PM3',qm_residues="917-919",
> /
> 1-916 is protein, 917 and 918 are two Magnesium ions, and 919 is ATP.
>
> The binding free energy has a large fluctuation, and some part of energy
> output results is shown:
>
> Energy Component Average Std. Dev. Std. Err. of Mean
>
> -------------------------------------------------------------------------------
> EGB -910.8454 711.7355
> 159.1489
> ESCF 199.0339 829.5371
> 185.4901
>
> DELTA G gas 607.9983 5.3856
> 1.2043
> DELTA G solv -917.4764 711.7595
> 159.1543
>
> DELTA TOTAL -110.4443 135.1893 30.2292
>
>
> Is there any wrong parameter setting in my input file for QM/MM-GBSA
> method? Would you like to give me some suggestions about this or QM/MM-GBSA
> calculation? Thank you!
>
First, including explicit ions in an MM/PBSA calculation is really
uncharted territory, so there is no guarantee that it will work...
The reason I think you are getting large variations is because you are
treating a single ion in addition to a ligand molecule as a single ligand.
If these two species bind independently or even semi-independently (i.e.,
they have distinct binding pockets and do not coordinate inside the
protein), then you can get very large fluctuations in the polar solvation
free energy (and electrostatic energy) depending on how close together
those the ligand and ion are (if the distance fluctuates a lot, so too will
the energy components).
Depending on the energy terms you want, you may have to re-think how you
define the 'receptor' and 'ligand'. It also could be the case that MM/PBSA
is an inappropriate method to use here...
Good luck,
Jason
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Apr 11 2013 - 07:30:03 PDT
