[gmx-users] correct rlist and Verlet scheme

Mark Abraham mark.j.abraham at gmail.com
Mon Feb 15 14:15:19 CET 2016


Hi,

On Mon, Feb 15, 2016 at 12:17 PM Timofey Tyugashev <tyugashev at niboch.nsc.ru>
wrote:

> I've studied the relevant sections of the manual, but I don't consider
> myself to be familiar enough with this field to successfully guess the
> right settings.
>
> ff99sb-ildn is included in the gromacs distribution, so shouldn’t be
> there some recommended settings for it?


Ideally, yes. But nobody has made a particular effort for that combination.

Or else how was it tested to run
> properly?
>

In principle, one would have to e.g. show that
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2970904/ can be replicated.
That's not a straightforward proposition...


> For now I've settled on the following parameters (the following is
> copied from mdout.mdp):
>
> ; NEIGHBORSEARCHING PARAMETERS
> ; cut-off scheme (Verlet: particle based cut-offs, group: using charge
> groups)
> cutoff-scheme            = Verlet
> ; nblist update frequency
> nstlist                  = 40
> ; ns algorithm (simple or grid)
> ns_type                  = grid
> ; Periodic boundary conditions: xyz, no, xy
> pbc                      = xyz
> periodic-molecules       = no
> ; Allowed energy error due to the Verlet buffer in kJ/mol/ps per atom,
> ; a value of -1 means: use rlist
> verlet-buffer-tolerance  = 0.005
> ; nblist cut-off
> rlist                    = 1.0
> ; long-range cut-off for switched potentials
> rlistlong                = 1.05
> nstcalclr                = -1
>

You've no doubt read
http://manual.gromacs.org/documentation/5.1.2/user-guide/mdp-options.html and
the comments above and realize that your choices for rlist and rlistlong
are ignored with the Verlet scheme. If you want to exactly follow the
paper, you need to set verlet-buffer-tolerance=-1, nstlist to whatever
implements 10fs given your choice of time step, and rlist to 1.05. This
will run rather slower than the Verlet-scheme defaults, particularly on
GPUs, for no known advantage. You can observe the difference in speed
yourself, of course. Your call.

Mark

; OPTIONS FOR ELECTROSTATICS AND VDW
> ; Method for doing electrostatics
> coulombtype              = PME
> coulomb-modifier         = Potential-shift-Verlet
> rcoulomb-switch          = 0
> rcoulomb                 = 1.0
> ; Relative dielectric constant for the medium and the reaction field
> epsilon-r                = 1
> epsilon-rf               = 0
> ; Method for doing Van der Waals
> vdw-type                 = Cut-off
> vdw-modifier             = Potential-shift-Verlet
> ; cut-off lengths
> rvdw-switch              = 0
> rvdw                     = 1.0
> ; Apply long range dispersion corrections for Energy and Pressure
> DispCorr                 = EnerPres
> ; Extension of the potential lookup tables beyond the cut-off
> table-extension          = 1
> ; Separate tables between energy group pairs
> energygrp-table          =
> ; Spacing for the PME/PPPM FFT grid
> fourierspacing           = 0.16
> ; FFT grid size, when a value is 0 fourierspacing will be used
> fourier-nx               = 0
> fourier-ny               = 0
> fourier-nz               = 0
> ; EWALD/PME/PPPM parameters
> pme_order                = 4
> ewald-rtol               = 1e-05
> ewald-rtol-lj            = 0.001
> lj-pme-comb-rule         = Geometric
> ewald-geometry           = 3d
> epsilon-surface          = 0
>
> And this is what the paper on it says:
>
>     Each system was equilibrated at 300 K and 1 atm with 2.4 ns of MD
>     simulation in the NPT ensemble. Then, MD simulations were carried
>     out in the NVT ensemble for 720 ns using the Nosé-Hoover thermostat
>     with a relaxation time of 1 ps. All simulations were performed using
>     the Desmond MD program^16
>     <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2970904/#b16> version
>     2.1.1.0 and either the Amber ff99SB^7
>     <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2970904/#b7> or the
>     modified Amber ff99SB force field described herein, which we have
>     termed ff99SB-ILDN. All bonds involving hydrogen atoms were
>     constrained with the SHAKE algorithm.^17
>     <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2970904/#b17> A cutoff
>     of 10 Å was used for the Lennard-Jones interaction and the
>     short-range electrostatic interactions. The smooth particle mesh
>     Ewald method^18
>     <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2970904/#b18> with a
>     32 × 32 × 32 grid and a fourth-order interpolation scheme was used
>     to compute the long-range electrostatic interactions. The pairlists
>     were updated every 10 fs with a cutoff of 10.5 Å. We used a
>     multistep RESPA scheme^19
>     <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2970904/#b19> for the
>     integration of the equations of motion with timesteps of 2.0, 2.0,
>     and 6.0 fs for the bonded, short-range nonbonded, and long-range
>     nonbonded interactions, respectively. To check for potential biases
>     introduced by long-range interactions between peptides in periodic
>     images, we repeated these simulations for four of the amino acids
>     (Xaa: Ile, Leu, Asp, and Asn) using a larger box with side length 37
>     Å. We found that the results of these control simulations were
>     within error of those using the smaller box sizes.
>
>
> 12.02.2016 22:06, gromacs.org_gmx-users-request at maillist.sys.kth.se пишет:
> > Message: 1
> > Date: Fri, 12 Feb 2016 14:09:28 +0000
> > From: Mark Abraham<mark.j.abraham at gmail.com>
> > To:"gromacs.org_gmx-users at maillist.sys.kth.se"
> >       <gromacs.org_gmx-users at maillist.sys.kth.se>,    "
> gmx-users at gromacs.org"
> >       <gmx-users at gromacs.org>
> > Subject: Re: [gmx-users] correct rlist and Verlet scheme
> > Message-ID:
> >       <
> CAMNuMAR1-sQgc4_tMWJtzv1UWy6+nKbwXaYZ2h+PAbA3Y-KjdQ at mail.gmail.com>
> > Content-Type: text/plain; charset=UTF-8
> >
> > Hi,
> >
> > As Adam says, you want to understand the reason why the forcefield
> > specifies the settings it does. For example, until at least AMBER 2003,
> > their force field papers did not specify PME parameters more precisely
> than
> > "x nm cutoff was used." If PME was used at all.;-)  So you can be guided
> > also by the parameters reported by successful users of these forcefields,
> > particularly those of the various force field and software authors.
> >
> > rlist is a free parameter so long as you choose it large enough to avoid
> > missing a significant number of interactions over the lifetime of the
> list
> > - but it affects performance strongly. rcoulomb can be varied for
> > equivalent accuracy if matching changes occur to PME parameters. VDW
> > parameters should generally follow the force field derivation precisely
> > (but add long-range corrections for dispersion) - GROMACS will not change
> > these. You should check out the background information in the reference
> > manual, sections 3.4.2 and 3.17.5
> >
> > Mark
> >
> > On Fri, Feb 12, 2016 at 1:56 PM Hardy, Adam<ah259 at hw.ac.uk>  wrote:
> >
> >> >There are people more qualified than me to explain the fine details of
> >> >this, but in brief:
> >> >
> >> >Verlet on it's own doesn't adjust rvdw or rcoloumb (rcoulomb may be
> >> >adjusted if you rae using PME for the coulomb interactions), but by
> default
> >> >it will adjust the neighbour list cut-off (rlist) to remain with a
> >> >specified energy tolerance given by verlet-buffer-tolerance. You can
> >> >overide this behaviour and manually set rlist by setting
> >> >verlet-buffer-tolerance=-1. Whether you want to or not will depend on
> why
> >> >the forcefield has specified that cut-off, it may well be to achieve
> the
> >> >same purpose, which is to propely catch interactions between atoms
> which
> >> >move in or out of the cut-off radii between rlist update steps.
> >> >
> >> >Adam
> >> >
> >> >________________________________________
> >> >From:gromacs.org_gmx-users-bounces at maillist.sys.kth.se  <
> >> >gromacs.org_gmx-users-bounces at maillist.sys.kth.se> on behalf of
> Timofey
> >> >Tyugashev<tyugashev at niboch.nsc.ru>
> >> >Sent: 12 February 2016 12:47
> >> >To:gromacs.org_gmx-users at maillist.sys.kth.se
> >> >Subject: [gmx-users] correct rlist and Verlet scheme
> >> >
> >> >I have trouble in correctly setting cut-offs with Verlet scheme to
> match
> >> >force-field specific settings.
> >> >It automatically adjusts cut-off and it looks like they change their
> >> >values from ones recommended for the force field.
> >> >Is there a way to correct it and set up them properly?
> >> >
> >> >The paper on FF lists 1.0 nm for short-range electrostatic and LJ
> >> >interactions.
> >> >The cut-off of pairlists is 1.05 nm.
> >> >
> >> >The forcefield is 99SB-ILDN and is present in GROMACS itself.
> >> >--
> >> >Gromacs Users mailing list
> >> >
> >> >* Please search the archive at
> >> >http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List  before
> >> >posting!
> >> >
> >> >* Can't post? Readhttp://www.gromacs.org/Support/Mailing_Lists
> >> >
> >> >* For (un)subscribe requests visit
> >> >https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users  or
> >> >send a mail togmx-users-request at gromacs.org.
> >> >--
> >> >Gromacs Users mailing list
> >> >
> >> >* Please search the archive at
> >> >http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List  before
> >> >posting!
> >> >
> >> >* Can't post? Readhttp://www.gromacs.org/Support/Mailing_Lists
> >> >
> >> >* For (un)subscribe requests visit
> >> >https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users  or
> >> >send a mail togmx-users-request at gromacs.org.
> >> >
>
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.


More information about the gromacs.org_gmx-users mailing list