Re: [AMBER] ntf parameters in energy minimization

From: Qin Zhang <zqyork.gmail.com>
Date: Tue, 30 Aug 2011 14:19:02 -0500

Dear Jason;
  Thank you very much.

On Mon, Aug 29, 2011 at 2:45 PM, Jason Swails <jason.swails.gmail.com>wrote:

> On Mon, Aug 29, 2011 at 1:24 PM, Qin Zhang <zqyork.gmail.com> wrote:
>
> > Dear All Amber users;
> > I would like to take an energy minimization only based on GB energy,
> > even though the energy model is not right. So I select ntf=8.
> > But it still reports VDWAALS and EEL energy.
> >
>
> That's because VDW, EEL, and EGB are all non-bonded interactions. You can
> zero out the VDW interactions by hacking the prmtop and setting every
> single
> ACOEF and BCOEF term equal to zero. However, decoupling the EEL and EGB
> calculations is impossible without hacking the code. However, I wouldn't
> exactly consider EEL and EGB to be separate terms. You can arrange the GB
> energy equation as a sum of terms in which one term takes the form of
> gas-phase electrostatic interactions. This is how Amber "decomposes" the
> electrostatic energy into an EEL term and an EGB term, but make sure you
> know exactly what you want here.
>
Yes. VDW, EEL and EGB are non-bonded interactions. I would like treat EGB as
solvation energy.
As you said, I can make VDW invalid by set ACOEFF and BCOEFF zeroes. For EEL
and EGB,
we need to modify the code. I just go into force.f in /src/sander/ of
Amber11. I just add several lines
 from 1008-1015 with one line comment.

   ! -- when igb!=0 and igb!=10, all nonbonds are done in routine egb:

      !zq added for test
      if ( ntf < 8 ) then
      pot%vdw = evdw
      pot%elec = eelt
      end if
      pot%gb = epol
      pot%surf = esurf
      pot%dvdl = dvdl

I just add from zq added for test and following 4 lines. It seems give me
the answer I need as

   NSTEP ENERGY RMS GMAX NAME NUMBER
      1 -2.5101E+03 4.6321E+00 4.2128E+01 NH2 1969

 BOND = 0.0000 ANGLE = 0.0000 DIHED =
0.0000
 VDWAALS = 0.0000 EEL = 0.0000 EGB =
-2510.0640
 1-4 VDW = 0.0000 1-4 EEL = 0.0000 RESTRAINT =
0.0000


   NSTEP ENERGY RMS GMAX NAME NUMBER
      2 -2.5106E+03 4.6282E+00 4.1761E+01 NH2 1969

 BOND = 0.0000 ANGLE = 0.0000 DIHED =
0.0000
 VDWAALS = 0.0000 EEL = 0.0000 EGB =
-2510.5923
 1-4 VDW = 0.0000 1-4 EEL = 0.0000 RESTRAINT =
0.0000


   NSTEP ENERGY RMS GMAX NAME NUMBER
      3 -2.5112E+03 4.6237E+00 4.1333E+01 NH2 1969

 BOND = 0.0000 ANGLE = 0.0000 DIHED =
0.0000
 VDWAALS = 0.0000 EEL = 0.0000 EGB =
-2511.2293
 1-4 VDW = 0.0000 1-4 EEL = 0.0000 RESTRAINT =
0.0000


   NSTEP ENERGY RMS GMAX NAME NUMBER
      4 -2.5120E+03 4.6186E+00 4.0834E+01 NH2 1969

 BOND = 0.0000 ANGLE = 0.0000 DIHED =
0.0000
 VDWAALS = 0.0000 EEL = 0.0000 EGB =
-2511.9983
 1-4 VDW = 0.0000 1-4 EEL = 0.0000 RESTRAINT =
0.0000


   NSTEP ENERGY RMS GMAX NAME NUMBER
      5 -2.5129E+03 4.6126E+00 4.0372E+01 OE1 1807

 BOND = 0.0000 ANGLE = 0.0000 DIHED =
0.0000
 VDWAALS = 0.0000 EEL = 0.0000 EGB =
-2512.9274
 1-4 VDW = 0.0000 1-4 EEL = 0.0000 RESTRAINT =
0.0000

But I am not sure whether my modification is right and also I am not sure
whether I can just paste several lines
on this forum for copyright issue even though we have valid license. If it's
not valid, please guide.




It give me the following results

>
> > NSTEP ENERGY RMS GMAX NAME NUMBER
> > 1 -1.2636E+04 4.6321E+00 4.2128E+01 NH2 1969
> >
> > BOND = 0.0000 ANGLE = 0.0000 DIHED =
> > 0.0000
> > VDWAALS = -1154.8397 EEL = -8971.0206 EGB =
> > -2510.0640
> > 1-4 VDW = 0.0000 1-4 EEL = 0.0000 RESTRAINT =
> > 0.0000
> >
> > when I tried ntf=7. The results are exactly the same. When I check the
> code
> > force.f in /src/sander, there seems no ntf<8 choice. Could you please
> > point
> > to me whether I can only select GB force for energy minimization? Or you
> > can
> > paste a right input file?
> > My input is as follows.
> >
> > GB energy minimization
> > &cntrl imin=1, maxcyc=10
> > cut=300.0, igb=1, saltcon=0, gbsa=0, extdiel=80.0, intdiel=1.0
> > ntpr=1, ntx=1, ntb=0, ntf=8
> > &end
> > &end
> >
>
> You do not need the second &end here. Also, I'm curious about why you do
> not set ntc=3 in this case. If you don't evaluate any internal potential
> terms, why aren't you constraining the length of every bond in the system?
> It would seem to me that you would need to in order to preserve the
> structure of your system and avoid getting nonsense (but I've never tried
> it
> so I don't know).
>
> Here, I know our model is definitely wrong and meaningless. We just want to
test GB results
so I didn't set ntc=3.

Thank you very much for your reply and suggestions.


> HTH,
> 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
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Aug 30 2011 - 12:30:04 PDT
Custom Search