RE: [AMBER] dftb_disper problem

From: <fmgarcia.cbm.uam.es>
Date: Fri, 8 May 2009 17:27:30 +0100

Hi Ross:
I'll try what you propose me. I'll keep you notified.
Thanks for all.

Fernando

> Hi Fernando,
>
> It is not based on atom type, simply element and the number of neighbors.
> In your case with H3O+ you may actually be triggering the 4th
> hybridization state every now and then. The same might be true of the
> hydrogens, you may end up with 3 atoms close to one of the hydrogens.
>
> Thus you may need to modify the file to read:
>
> O 0.560 0.560 0.560 0.560 3.8 3.8 3.8 3.8 3.15
> H 0.386 0.386 0.386 0.000 3.5 3.5 3.5 3.5 0.8
>
> Of course be aware that this is effectively 'guessing' the parameters here
> so just keep that in mind when interpreting any results.
>
> Also make sure the code picks up the new file when it runs. The dispersion
> data should be written to the output file.
>
> Good luck,
> Ross
>
>> -----Original Message-----
>> From: amber-bounces.ambermd.org [mailto:amber-bounces.ambermd.org] On
>> Behalf Of Fernando Martín García
>> Sent: Friday, May 08, 2009 5:43 AM
>> To: AMBER Mailing List
>> Subject: Re: [AMBER] dftb_disper problem
>>
>> Hi Gustavo:
>> I have changed the parameter that you indicated to me, but the problem
>> persist. I was investigating which is the atom that provokes the
>> problems and it's a oxygen atom that belongs to a water molecule, that
>> needs to be protonated (H3O+)at the end of the reaction I'm simulating.
>> These values depend on the type of oxygen atom I'm using?
>>
>> Thanks
>> Fernando
>>
>> El jue, 07-05-2009 a las 09:01 -0400, Gustavo Seabra escribió:
>> > Hi Fernando,
>> >
>> > This is actually a common problem with the use of dispersion in DFTB.
>> > What is happening here is that, at every step, the program tries to
>> > guess a hibridization state for an atom by counting the number of
>> > atoms within a certain distance threshold, in this case, 1.2A. Then,
>> > from the DISPERSION.INP_ONCHSP, it chooses the polarizability data
>> for
>> > this hibridization. The DISPERSION.INP_ONCHSP file has the following
>> > format:
>> >
>> > $ cat DISPERSION.INP_ONCHSP
>> > O 0.560 0.560 0.000 0.000 3.8 3.8 3.8 3.8 3.15
>> > N 1.03 1.03 1.090 1.090 3.8 3.8 3.8 3.8 2.82
>> > C 1.382 1.382 1.382 1.064 3.8 3.8 3.8 3.8 2.5
>> > H 0.386 0.386 0.000 0.000 3.5 3.5 3.5 3.5 0.8
>> > P 1.6 1.6 1.6 1.6 4.7 4.7 4.7 4.7 4.5 !PO4
>> > S 3.0 3.0 3.0 3.0 4.7 4.7 4.7 4.7 4.8 !S not SO2
>> >
>> > The first 4 numbers after the atom name are the polarizabilities for
>> > each hibridization state, defined by the number of neighbors. (To
>> > understand the parameterization, look at the original reference: M.
>> > Elstner, P. Hobza, T. Frauenheim, S. Suhai, and E. Kaxiras, J. Chem.
>> > Phys. vol. 114 (2001) 5149 ) Notice that those parameters have been
>> > derived for DNA bases, so there is no guarantee they will be
>> > appropriate for anything else.
>> >
>> > Basically, these parameters were derived depending on the number of
>> > neighbors you may expect to find for an atom. However, during an MD
>> > calculation, it eventually happens that an extra atom gets close
>> > enough to confuse the program. In your case you have an Oxygen (atom
>> > 2) with 3 neighbors, and there is no polarizability data in the file
>> > for this situation. in this case, sander complains and crashes.
>> >
>> > ============
>> > WORKAROUND
>> > ============
>> > In the Oxygen line, just copy the parameter for 2 neighbors in to the
>> > next column. it will not be precise, but it will preventing sander
>> > from crashing. Line 1 in the DISPERSION.INP_ONCHSP file will will
>> then
>> > look like:
>> > O 0.560 0.560 0.560 0.000 3.8 3.8 3.8 3.8 3.15
>> >
>> > Good luck,
>> > Gustavo Seabra
>> > Postdoctoral Associate
>> > Quantum Theory Project - University of Florida
>> > Gainesville - Florida - USA
>> >
>> >
>> >
>> > 2009/5/7 Fernando Martín García <fmgarcia.cbm.uam.es>:
>> > > Hi all:
>> > >
>> > > I have a problem with the flag "dftb_disper". When i launch my job
>> with
>> > > a value of dftb_disper= 0, the job run well, but if i try to use
>> > > dftb_disper =1, the following message appears:
>> > >
>> > > -------------------------------------------------------------------
>> -------------
>> > > 4. RESULTS
>> > > -------------------------------------------------------------------
>> -------------
>> > >
>> > > *******************************************
>> > > WARNING: a parameter is 0.0 for atom 2
>> > > IZP = 2
>> > > nei = 4 ( 3 neighbors.)
>> > > hh1 = 0.00
>> > > hh2 = 3.80
>> > > Ni = 3.15
>> > > *******************************************
>> > > SANDER BOMB in subroutine <dispersion_params>
>> > > qm2_dftb_dispersion_params.f
>> > > Exiting.
>> > >
>> > > These are the parameters that i use in this job:
>> > >
>> > > &cntrl
>> > > ntx = 7, irest = 1, ntrx = 1,
>> > > ntxo = 1, nmropt = 1, ntr = 0,
>> > > ntpr = 100, ntwx = 100,
>> > > ntf = 1, ntb = 2, dielc = 1.0,
>> > > cut = 9, nsnb = 10, scnb = 2.0, scee = 1.2,
>> > > imin = 0, ibelly=0,
>> > > nstlim = 1000, dt = 0.001,
>> > > temp0 = 300.0, tempi = 300.0,
>> > > ntt = 1, vlimit = 20.0,
>> > > ntp = 1,
>> > > ntc = 1, tol = 0.00001, pres0=1, comp=44.6,
>> > > jfastw=0, nscm=1000,
>> > > ifqnt=1
>> > > /
>> > > &qmmm
>> > > iqmatoms= 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
>> 17,
>> > > 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28,
>> > > qmcharge= 0,
>> > > qm_theory='DFTB',
>> > > dftb_disper= 1,
>> > > dftb_maxiter= 70,
>> > > qmshake= 0,
>> > > qmmm_int = 2,
>> > > peptide_corr = 0,
>> > > qmcut= 8.0
>> > >
>> > > I've downloaded the parameter file "mio" from www.dftb.org and
>> installed
>> > > in $AMBERHOME/dat/slko, and also the other two files are in the
>> same
>> > > directory (DISPERSION.INP_ONCHSP and CM3). Can anybody tell me
>> where is
>> > > the problem?
>> > >
>> > > Thanks
>> > >
>> > > --
>> > > Fernando Martín García.
>> > > Centro de Biología Molecular "Severo Ochoa".
>> > > C/ Nicolás Cabrera, 1.
>> > > Campus UAM. Cantoblanco, 28049 Madrid. Spain.
>> > >
>> > >
>> > > _______________________________________________
>> > > 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
>> --
>> Fernando Martín García.
>> Centro de Biología Molecular "Severo Ochoa".
>> C/ Nicolás Cabrera, 1.
>> Campus UAM. Cantoblanco, 28049 Madrid. Spain.
>>
>>
>> _______________________________________________
>> 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
>


Fernando Martín García.
Centro de Biología Molecular "Severo Ochoa".
C/ Nicolás Cabrera, 1.
Campus UAM. Cantoblanco, 28049 Madrid. Spain.


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed May 20 2009 - 15:12:14 PDT
Custom Search