AMBER Archive (2007)

Subject: Re: AMBER: solvateCap minin and mdin for qm-mm dftb

From: Francesco Pietra (chiendarret_at_yahoo.com)
Date: Fri Aug 24 2007 - 11:01:31 CDT

Hi Bud and Ross:
Thanks

When launching the preliminary minimization with adjusted

Initial min for qmmm with dftb
&cntrl
imin=1, maxcyc=500, ncyc=200,
cut=20.0, ntb=0, ntc=2, ntf=2,
ivcap=0, fcap=10, igb=10,
ifqnt=1
/
&qmmm
qmmask=':1',
qmcharge=0,
qmtheory=7,
qmshake=0,
qm_ewald=0
/

sander bombed. min_qmmm_dftb.out ended with:

LOADING THE QUANTUM ATOMS AS GROUPS
     Mask :1; matches 98 atoms
 SANDER BOMB in subroutine read_qmmm_namelist
 qmtheory = DFTB but qmgb > 1. DFTB does not currently
 support full QM GB. Only qmgb values of 0 or 1 are supported.

Once flag "qmgb=0" is added just after "qmtheory=7", the minimization is going
on. From tail -f it is at NSTEP 50.

I was unable to detect where GB is stated in the minin. I also wonder about the
difference between "qmgb=0" and qmgb=1" which I did not fin in either Amber9
manual or the pdf included in Marcus Elnster files.

If there is nothing wrong about GB statements, I assume that qmgb=0 (or qmgb=1)
flag should also be added to mdin, modified (for a trial) as follows

300K constant temp for qmmd with dftb
 &cntrl
  imin=0, ntb=0
  cut=20.0, ntc=2, ntf=2,
  tempi=300.0, temp0=300.0,
  ntt=3, gamma_ln=1.0,
  nstlim=1000, dt=0.002,
  ivcap=0, fcap=10, igb=10
  ntpr=10, ntwx=10,ifqnt=1
 /
 &qmmm
  qmmask=':1',
  qmcharge=0,
  qmtheory=7,
qmgb=0,
  qmshake=0,
  qm_ewald=0
 /

Thanks

francesco

--- "M. L. Dodson" <mldodson_at_houston.rr.com> wrote:

> Francesco Pietra wrote:
> > I am trying to build prmtop and inpcrd for qm-mm with theory=7 (dftb). I am
> > stuck at solvateCap, argument #3 (position).
> >
> > Using *.prepin and *frcmod from previous successful runs with theory=2,
> >
> > startx
> >
> > $AMBERHOME/exe/xleap -s -f $AMBERHOME/dat/leap/cmod/leaprc.ff99
> >
> > source leaprc.gaff
> >
> > loadamberprep EQE.prepin
> >
> > loadamberparams EQE.frcmod
> >
> > check EQE (unit is OK)
> >
> > solvateCap EQE TIP3PBOX EQE 20.0
> >
> > returns
> >
> > Argument #3 is type Unit must be of type: [molecule residue atom list]
> >
> > Using as Argument #3: "1" (residue number in prepin), "C17" (a central
> carbon
> > atom number in prepin, "c" the C17 atom type, or, finally (and blindly)
> > EQE.2.CA the error is: Argument #3 is type String.
> >
> > I am short of imagination (even consulting the manual and the mailing list)
> of
> > what should be given in my case as Argument #3.
>
> Without getting into the structured design of LEaP, perhaps the
> easiest way to for you to proceed is to pick the atom you want to
> make the center of your spherical cap. For purposes of
> illustration, let us assume the Cartesian coordinates of that
> atom are:
>
> x = 1.0, y = 2.0, z = 3.0
>
> Then a solvate cap command using these coordinates might be:
>
> solvateCap EQE TIP3PBOX {1.0 2.0 3.0} 20.0
>
> The desc command is your friend when trying to learn the syntax of
> LEaP. You just have to practice.
>
> > ________________--
> > That said, on summing up info from the manual and suggestions from Bud and
> > Gustavo, I have a number of doubts about min.in and md.in, in particular
> if
> > qm_pme should be omitted at all. The two files I have set up read:
> >
> > Initial min for qmmm with dftb
> > &cntrl
> > imin=1, maxcyc=500, ncyc=200,
> > cut=20.0, ntb=1, ntc=2, ntf=2,
> > ivcap=0, fcap=10
> > ifqnt=1
> > /
> > &qmmm
> > qmmask=':1',
> > qmcharge=0,
> > qmtheory=7,
> > qmshake=0,
> > qm_ewald=0
> > /
> >
> > 300K constant temp for qmmd with dftb
> > &cntrl
> > imin=0, ntb=1
> > cut=20.0, ntc=2, ntf=2,
> > tempi=300.0, temp0=300.0,
> > ntt=3, gamma_ln=1.0,
> > nstlim=20000, dt=0.002,
> > ivcap=0, fcap=10,
> > ntpr=100, ntwx=100,ifqnt=1
> > /
> > &qmmm
> > qmmask=':1',
> > qmcharge=0,
> > qmtheory=7,
> > qmshake=0,
> > qm_ewald=0
> > /
> >
> >
> > I want to try if, at the current implementation of qm-mmm/dftb
> > in Amber9, speed is acceptable for a 98-atoms molecule. If
> > acceptable, to compare ESCF and EPtot with AM1.
> >
>
> Since amber9 sander cannot do pme with QMMM and qmtheory=7, you
> should not turn on qm_pme. Also, for solvent cap simulations,
> ntb=0, not ntb=1 as you have specified. And you should think
> about the suggestion on page 99 of the manual to set igb=10. I
> also think you may have to reduce the timestep. I use 0.5fs for
> QMMM steered MD, but you may be able to get by with 1.0fs, 2.0fs
> is generally workable, but is said to be borderline for classical
> MD. See the email list archives.
>
> Of course the definition of acceptable speed is personal, but I
> have found 90+ atom QM regions to be workable with qmtheory=7.
>
> >
> >
> > Thanks for checking
> >
> > francesco pietra
>
> HTH,
>
> Bud Dodson
> --
> M. L. Dodson
> Email: mldodson-at-houston-dot-rr-dot-com
> Phone: eight_three_two-five_63-386_one
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>

       
____________________________________________________________________________________
Be a better Globetrotter. Get better travel answers from someone who knows. Yahoo! Answers - Check it out.
http://answers.yahoo.com/dir/?link=list&sid=396545469
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu