Re: [gmx-users] 1-4 interaction not within cut-off

[Justin A. Lemkul]

Rabab Toubar wrote:
they are stating explicitly that it is a twin range cutoff
the manual states that
"vdwtype:
     Cut-off
          Twin range cut-off’s with neighbor list cut-off rlist and VdW 
cut-off rvdw, where
          rvdw ≥ rlist."
"coulombtype:
     Cut-off
          Twin range cut-off’s with neighborlist cut-off rlist and 
Coulomb cut-off rcoulomb, where rcoulomb ≥ rlist."

because rlist is not stated so I left it as default value 1nm according 
to the manual because rcoulomb and rvdw can both be ≥ rlist.

These are generalizations.  Since a plain cutoff is essentially a special case 
of a twin-range cutoff, then it can be stated that cutoffs are "twin range."

If rlist=rcoulomb=rvdw you most certainly do not have a twin-range nonbonded 
setup.  It is a plain truncation.  Please consult some basic textbook material 
for the theory and proper implementation of a twin-range cutoff scheme.


the authors didn't set any constraints, and I didn't use any. For the 
restraints I used pr.mdp and this step went well.


They most certainly did set constraints!  From the text you quoted:

"...time step for all simulations is 2fs and LINCS method was used to constrain 
bonds lengths."

Is there any changes I can do for the md.mdp while sticking to the 
paper's methods, like may be increasing the table-extension as it was 
suggested in the error message?

Never adjust the table-extension parameter.  The simple fact is that your system 
is blowing up, and no amount of coaxing this parameter into cooperation will fix 
broken model physics.  I see two possibilities for fixing the instability:

1. Use constraints with a 2-fs timestep
2. Do not use a plain Coulombic cutoff.  Use PME.  I understand you're trying to 
follow a previously-described methodology, but the text you quoted simply 
neglects long-range interactions, both LJ and Coulombic.  Twin-range cutoffs 
apply to short-range interactions, and do not preclude the use of a stable 
long-range method like PME.


And if I want to define the electrostatics properly, is there anything 
more than setting the coulombtype for electrostatics as PME? or this 
would be only for short range electrostatics and another term is used 
for long range elecrtrostatics?


PME is a long-range method, dealing with terms that fall beyond rcoulomb. 
Anything within rcoulomb is short-range.

I know I won't learn GMX in a few days, but I am trying to get all the 
help that I could get


If you're just starting out, then textbook reading and tutorials are absolutely 
requisite.  I would also recommend trying to find a paper with more detailed 
methodology.  IMHO, the text you quote leaves out several important details, 
making the results difficult to reproduce.

-Justin

Thanks
Rabab Toubar


--- On *Fri, 6/4/10, Justin A. Lemkul /<jalem...@vt.edu>/* wrote:


    From: Justin A. Lemkul <jalem...@vt.edu>
    Subject: Re: [gmx-users] 1-4 interaction not within cut-off
    To: "Gromacs Users' List" <gmx-users@gromacs.org>
    Date: Friday, June 4, 2010, 4:49 PM



    Rabab Toubar wrote:
     > Well all the relevant info from m&m is attached below. I tried to
    contact the author with no reply!. Ref: Journal of Biomolecular
    Structure & Dynamics, ISSN 0739-1102
     > Volume 27, Issue Number 4, (2010)
     >
     > "The GROMACS 3.3.3 simulation package (www.gromacs.org) was used
    for all the simulations with ffG45a3 force field. The SPC water
    model was used in the simulations. Periodic boundary conditions were
    applied to the system. A twin range cut-off of 1.0 nm for van der
    Waals and Coulomb interactions were used."

    This is not fully explained.  The value of rlist is not given, but
    based on this information, I would not conclude that this is a
    twin-range cutoff - the values are the same, so there are not two
    interaction ranges!

     > "All simulations were performed at a constant temperature of 300
    K using Berendson barostat with isotropic coupling constant of τp 5
    0.5 ps,  compressibility equal to 4.5 3 10-5 bar-1 and time
    constants equal to 0.1 ps for the three groups i.e., protein,
    solvent and Cl-. Integration

    I would not recommend following this part of the protocol.  Coupling
    ions separately from the solvent often leads to instability:

    http://www.gromacs.org/Documentation/Terminology/Thermostats

     > time step for all simulations is 2fs and LINCS method was used to
    constrain bonds lengths. The structures were energy minimized for
    1000 steps using steepest descent prior to the MDS setup. Positional
    restraints were applied to the protein (all-bonds) with a force
    constant of 1000 kJ mol-1 for 2 ns to enable the relaxation of water
    molecules around the protein. All restraints were then removed, and
    the simulations were carried out. "

    Make sure that you understand that constraints and restraints are
    different things:

    http://www.gromacs.org/Documentation/Terminology/Constraints_and_Restraints

    By removing the restraints, it should not be inferred that the
    constraints were removed as well.

    The absence of any information about long-range electrostatics is
    also somewhat troubling.  A plain cutoff is not generally recommended.

    -Justin

     >
     >
     > Thanks
     > Rabab Toubar
     >
     >
     > --- On *Fri, 6/4/10, Justin A. Lemkul /<jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>>/* wrote:
     >
     >
     >     From: Justin A. Lemkul <jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>>
     >     Subject: Re: [gmx-users] 1-4 interaction not within cut-off
     >     To: "Gromacs Users' List" <gmx-users@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>
     >     Date: Friday, June 4, 2010, 3:51 PM
     >
     >
     >
     >     Rabab Toubar wrote:
     >      > Hi Justin
     >      >
     >      > Thanks a lot for your prompt reply.
     >      > The end of the em log file says:
     >      > Stepsize too small, or no change in energy.
     >      > Converged to machine precision,
     >      > but not to the requested precision Fmax < 10
     >      >
     >      > Double precision normally gives you higher accuracy.
     >      >
     >      > Steepest Descents converged to machine precision in 206 steps,
     >      > but did not reach the requested Fmax < 10.
     >      > Potential Energy  = -1.8784400e+05
     >      > Maximum force     =  7.9173694e+02 on atom 159
     >      > Norm of force     =  6.5868745e+03
     >      >
     >
     >     OK, that all looks normal.
     >
     >      > And after em, I ran pr for 2ns and it went all good and you're
     >     correct it was for all bonds.
     >      > After I read the manual and a couple of tutorial I understood
     >     that the PME is better for electrostatics but I am reproducing a
     >     reported work that used twin range cutoff of 1 for both vdw and
     >     electrostatics using 2fs time step and I want to see if I
    will get
     >     similar results. The authors stated that the simulations went
    well
     >     till the end of the 25 ns run! But seems that mine is
    crashing. Any
     >     suggestions?
     >      >
     >
     >     If you have all the short-range cutoffs set to an equal
    value, that
     >     is decidedly not a twin-range cutoff.  Even the use of a proper
     >     twin-range setup does not preclude the use of a long-range
     >     electrostatics method like PME.  I would suggest you look a
    bit more
     >     closely at the methods you're trying to reproduce. I also suspect
     >     that you'll need constraints in order to use a 2-fs timestep.
     >
     >     -Justin
     >
     >      > Thanks
     >      >
     >      > Rabab Toubar
     >      >
     >      >
     >      > --- On *Fri, 6/4/10, Justin A. Lemkul /<jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>
     >     </mc/compose?to=jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>>>/* wrote:
     >      >
     >      >
     >      >     From: Justin A. Lemkul <jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>
     >     </mc/compose?to=jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>>>
     >      >     Subject: Re: [gmx-users] 1-4 interaction not within
    cut-off
     >      >     To: "Gromacs Users' List" <gmx-users@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>
     >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>>
     >      >     Date: Friday, June 4, 2010, 3:25 PM
     >      >
     >      >
     >      >
     >      >     Rabab Toubar wrote:
     >      >      > Hi Justin,
     >      >      >
     >      >      > Thanks for the reply. When I tried to gmxcheck the
     >     minimization
     >      >     process I got an error:
     >      >      >
     >      >      > Timesteps at t=810 don't match (2, 1)
     >      >      >
     >      >      > Timesteps at t=814 don't match (1, 2)
     >      >      >
     >      >      > Timesteps at t=816 don't match (2, 1)
     >      >      >
     >      >      > Timesteps at t=818 don't match (1, 2)
     >      >      >
     >      >      > till the end of the run, that is why I thought the
    system
     >     is not
     >      >     properly minimized. I will look into how to check the
     >     Potential Energy.
     >      >      >
     >      >
     >      >     When running EM, mdrun prints the important
    information out
     >     at the
     >      >     very end, both to the screen and to the end of the
    .log file.
     >      >
     >      >
     >      >      > the em.mdp file is as follows:
     >      >      > cpp          = /usr/bin/cpp ; the c preprocessor
     >      >      > define       = -DFLEXIBLE
     >      >      > constraints  = none
     >      >      > integrator   = steep
     >      >      > dt           = 0.002  ;ps ie 2 fs!
     >      >      > nsteps       = 1000
     >      >      > nstlist      = 10       ns_type      = grid
     >      >      > rlist         = 1.0     ;short range
     >      >      > coulombtype  = cut-off
     >      >      > rcoulomb     = 1.0    rvdw         = 1.0
     >      >      > vdwtype      = cut-off  fourierspacing  = 0.12
     >      >      >  fourier_nx = 0
     >      >      >  fourier_ny = 0
     >      >      >  fourier_nz = 0
     >      >      >  pme_order = 4
     >      >      >  ewald_rtol = 1e-5
     >      >      >  optimize_fft = yes   emtol        = 10   emstep   
       = 0.01
     >      >      >
     >      >      > the md.mdp file is:
     >      >      > cpp                 = /usr/bin/cpp
     >      >      > constraints         =  none                 ;*
     >      >      > integrator          =  md
     >      >      > dt                  =  0.002  ; ps !
     >      >
     >      >     Using a 2-fs timestep without constraints may not be
    stable.     Your
     >      >     original message said you ran 2 ns of position
    restrained MD with
     >      >     LINCS, which I assumed meant that you were
    constraining the bond
     >      >     lengths.  If you actually did this, what reason do you
    have for
     >      >     turning off the restraints after equilibrating?
     >      >
     >      >      > nsteps              =   12500000    ; total 25 ns;*
     >      >      > nstcomm             =  1
     >      >      > nstxout             =  500    ; collect data every 1 ps
     >      >      > nstvout             =  0
     >      >      > nstfout             =  0
     >      >      > nstlist             =  10
     >      >      > ns_type             =  grid
     >      >      > rlist               =  1.0
     >      >      > coulombtype         =  cut-off
     >      >
     >      >     Using a plain cutoff for electrostatics can lead to a
    number of
     >      >     artefacts.  I see no reason to use this method in
    favor of other,
     >      >     more advanced and accurate methods.
     >      >
     >      >     -Justin
     >      >
     >      >      > rcoulomb            =  1.0
     >      >      > vdwtype             =  cut-off
     >      >      > rvdw                =  1.0            ;*
     >      >      > fourierspacing    =  0.12
     >      >      > fourier_nx        =  0
     >      >      > fourier_ny        =  0
     >      >      > fourier_nz        =  0
     >      >      > pme_order         =  4
     >      >      > ewald_rtol        =  1e-5
     >      >      > optimize_fft            =  yes          > ; Berendsen
     >     temperature coupling is on
     >      >      > Tcoupl                =  berendsen
     >      >      > tau_t                 =  0.1    0.1
     >      >      > tc-grps               =      protein    non-protein
    ref_t                       =  300        300
     >      >      >
     >      >      > ; Pressure coupling is on
     >      >      > Pcoupl              =  berendsen
     >      >      > pcoupltype        =  isotropic
     >      >      > tau_p               =  0.5    ; ps
     >      >      > compressibility     =  4.5e-5    ; bar-1
     >      >      > ref_p               =  1.0
     >      >      >
     >      >      > ; Generate velocites is on at 300 K.
     >      >      > gen_vel             =  yes
     >      >      > gen_temp            =  300.0
     >      >      > gen_seed            = 621432
     >      >      >
     >      >      > Any suggestions are appreciated
     >      >      >
     >      >      > Thanks
     >      >      > Rabab Toubar
     >      >      >
     >      >      >
     >      >      > --- On *Thu, 6/3/10, Justin A. Lemkul
    /<jalem...@vt.edu </mc/compose?to=jalem...@vt.edu>
     >     </mc/compose?to=jalem...@vt.edu </mc/compose?to=jalem...@vt.edu>>
     >      >     </mc/compose?to=jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>
     >     </mc/compose?to=jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>>>>/* wrote:
     >      >      >
     >      >      >
     >      >      >     From: Justin A. Lemkul <jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>
     >     </mc/compose?to=jalem...@vt.edu </mc/compose?to=jalem...@vt.edu>>
     >      >     </mc/compose?to=jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>
     >     </mc/compose?to=jalem...@vt.edu
    </mc/compose?to=jalem...@vt.edu>>>>
     >      >      >     Subject: Re: [gmx-users] 1-4 interaction not within
     >     cut-off
     >      >      >     To: "Discussion list for GROMACS users"
     >      >     <gmx-users@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>
    </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>
     >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>
     >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>>>
     >      >      >     Date: Thursday, June 3, 2010, 4:37 PM
     >      >      >
     >      >      >
     >      >      >
     >      >      >     Rabab Toubar wrote:
     >      >      >      > Hi,
     >      >      >      >
     >      >      >      > I set my .mdp file to minimize the system
    for 1000
     >     steps,
     >      >     but it
     >      >      >     stopped at ~200 saying: "Stepsize too small, or no
     >     change in
     >      >     energy.
     >      >      >      > Converged to machine precision, but not to the
     >     requested
     >      >      >     precision Fmax < 10" but the molecule is now
    outside
     >     the box
     >      >      >      >
     >      >      >
     >      >      >     This is not really a problem, provided that
    both the
     >     potential
     >      >      >     energy and Fmax that resulted are
     >      >      >       
       reasonable.http://www.gromacs.org/Documentation/Errors#Stepsize_too_small.2c_or_no_change_in_energy._Converged_to_machine_precision.2c_but_not_to_the_requested_precision
     >      >      >
     >      >      >      > I did position restraint for 2ns using
    LINCS, and
     >     things
     >      >     went well
     >      >      >      >
     >      >      >      > running md, it ran for only 0.3 ns out of
    30, and I
     >     got an
     >      >     error
     >      >      >     message that 1-4 interaction between 2 atoms
    are at a
     >     distance >
     >      >      >     table-size (1nm) with a suggestion to increase the
     >      >     table-extension
     >      >      >     in mdp file.
     >      >      >      > I checked       
       http://www.gromacs.org/Documentation/Errors#1-4_interaction_not_within_cut-off
     >      >      >     suggesting reminimizing.
     >      >      >      >
     >      >      >
     >      >      >     The other possibility is that your .mdp
    settings are
     >      >     inappropriate,
     >      >      >     but you haven't provided that information.
     >      >      >
     >      >      >      > My question is: is it OK that the system doesn't
     >     minimize till
     >      >      >     the end of the specified number of steps? is it
    OK to be
     >      >     outside the
     >      >      >     box as long as you assign boundary conditions as in
     >     case of NAMD?
     >      >      >     and what is the best option - you think - I have to
     >     solve the
     >      >     md problem
     >      >      >      >
     >      >      >
     >      >      >     Whether or not you need to revisit the EM procedure
     >     depends
     >      >     on how
     >      >      >     well it actually did (see comment above).  The
    periodicity
     >      >     effect is
     >      >      >     a non-issue, since there is no "outside" of a
    periodic
     >     box.
     >      >      >
     >      >      >       
       http://www.gromacs.org/Documentation/Terminology/Periodic_Boundary_Conditions
     >      >      >
     >      >      >     -Justin
     >      >      >
     >      >      >      > Any suggestions are highly apprecitaed
     >      >      >      > Rabab
     >      >      >      >
     >      >      >      >
     >      >      >
     >      >      >     -- ========================================
     >      >      >
     >      >      >     Justin A. Lemkul
     >      >      >     Ph.D. Candidate
     >      >      >     ICTAS Doctoral Scholar
     >      >      >     MILES-IGERT Trainee
     >      >      >     Department of Biochemistry
     >      >      >     Virginia Tech
     >      >      >     Blacksburg, VA
     >      >      >     jalemkul[at]vt.edu | (540) 231-9080
     >      >      > 
       http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
     >      >      >
     >      >      >     ========================================
     >      >      >     -- gmx-users mailing list   
    gmx-users@gromacs.org </mc/compose?to=gmx-us...@gromacs.org>
     >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>
     >      >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>
     >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>>
     >      >      >     http://lists.gromacs.org/mailman/listinfo/gmx-users
     >      >      >     Please search the archive at
     >     http://www.gromacs.org/search before
     >      >      >     posting!
     >      >      >     Please don't post (un)subscribe requests to the
    list.
     >     Use the www
     >      >      >     interface or send it to
    gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>
     >     </mc/compose?to=gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>>
     >      >     </mc/compose?to=gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>
     >     </mc/compose?to=gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>>>.
     >      >      >     Can't post? Read
     >     http://www.gromacs.org/mailing_lists/users.php
     >      >      >
     >      >      >
     >      >
     >      >     -- ========================================
     >      >
     >      >     Justin A. Lemkul
     >      >     Ph.D. Candidate
     >      >     ICTAS Doctoral Scholar
     >      >     MILES-IGERT Trainee
     >      >     Department of Biochemistry
     >      >     Virginia Tech
     >      >     Blacksburg, VA
     >      >     jalemkul[at]vt.edu | (540) 231-9080
     >      >     http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
     >      >
     >      >     ========================================
     >      >     -- gmx-users mailing list    gmx-users@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>
     >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>
     >      >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>
     >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>>
     >      >     http://lists.gromacs.org/mailman/listinfo/gmx-users
     >      >     Please search the archive at
    http://www.gromacs.org/search before
     >      >     posting!
     >      >     Please don't post (un)subscribe requests to the list.
    Use the www
     >      >     interface or send it to gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>
     >     </mc/compose?to=gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>>
     >      >     </mc/compose?to=gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>
     >     </mc/compose?to=gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>>>.
     >      >     Can't post? Read
    http://www.gromacs.org/mailing_lists/users.php
     >      >
     >      >
     >
     >     -- ========================================
     >
     >     Justin A. Lemkul
     >     Ph.D. Candidate
     >     ICTAS Doctoral Scholar
     >     MILES-IGERT Trainee
     >     Department of Biochemistry
     >     Virginia Tech
     >     Blacksburg, VA
     >     jalemkul[at]vt.edu | (540) 231-9080
     >     http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
     >
     >     ========================================
     >     -- gmx-users mailing list    gmx-users@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>
     >     </mc/compose?to=gmx-us...@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>>
     >     http://lists.gromacs.org/mailman/listinfo/gmx-users
     >     Please search the archive at http://www.gromacs.org/search before
     >     posting!
     >     Please don't post (un)subscribe requests to the list. Use the www
     >     interface or send it to gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>
     >     </mc/compose?to=gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>>.
     >     Can't post? Read http://www.gromacs.org/mailing_lists/users.php
     >
     >

    -- ========================================

    Justin A. Lemkul
    Ph.D. Candidate
    ICTAS Doctoral Scholar
    MILES-IGERT Trainee
    Department of Biochemistry
    Virginia Tech
    Blacksburg, VA
    jalemkul[at]vt.edu | (540) 231-9080
    http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

    ========================================
    -- gmx-users mailing list    gmx-users@gromacs.org
    </mc/compose?to=gmx-us...@gromacs.org>
    http://lists.gromacs.org/mailman/listinfo/gmx-users
    Please search the archive at http://www.gromacs.org/search before
    posting!
    Please don't post (un)subscribe requests to the list. Use the www
    interface or send it to gmx-users-requ...@gromacs.org
    </mc/compose?to=gmx-users-requ...@gromacs.org>.
    Can't post? Read http://www.gromacs.org/mailing_lists/users.php



--
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================
--
gmx-users mailing list    gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php