[gmx-users] What is the most reliable way to run repeats for reproducibility?

Mark Abraham mark.j.abraham at gmail.com
Wed Jan 10 14:58:44 CET 2018


Hi,

See http://www.gromacs.org/Documentation/Terminology/Reproducibility. Some
people think they want reproducibility of a trajectory, which is generally
not needed, and not consistent with highly efficient sampling (particularly
with GPUs involved).

But what you actually want is not reproducibility. The usual approach is to
change the random seed used when you first generate velocities, and rely on
the natural chaotic divergence during equilibration to lead to independent
sampling in the replicas.

Mark

On Wed, Jan 10, 2018 at 2:12 PM ZHANG Cheng <272699575 at qq.com> wrote:

> Dear Gromacs,
> I can think of different ways of running repeats, after reading Justin's
> lysozyme tutorial.
>
>
> The 1st way: all starting from the same em.tpr after energy minimization
> (EM) and use em.tpr individually for subsequent steps (NVT, NPT and
> production MD):
> ) repeat 1: same em.tpr → NVT → NPT → md_0_1.tpr→ production MD
> ) repeat 2: same em.tpr → NVT → NPT → md_0_1.tpr→ production MD
> ) repeat 3: same em.tpr → NVT → NPT → md_0_1.tpr→ production MD
> ......
>
>
> The 2nd way: all starting from the same md_0_1.tpr and use it for
> different production MD:
> ) repeat 1: same em.tpr → same NVT → same NPT → same md_0_1.tpr→
> production MD
> ) repeat 2: same md_0_1.tpr→ production MD
> ) repeat 3: same md_0_1.tpr→ production MD
> ......
>
>
>
> The 3rd way: all starting from the same check point file within the
> production run and use it for the rest of the production MD:
> ) repeat 1: same em.tpr → same NVT → same NPT → same md_0_1.tpr→ same
> production MD for 50 ns → same .cpt file → production MD for another 200 ns
> ) repeat 2: same .cpt file → production MD for another 200 ns
> ) repeat 3: same .cpt file → production MD for another 200 ns
> ......
>
>
>
> Of course, the 3rd way is easier. But does it mean it may not cover enough
> conformations, as they tend to be more resembled from each other than the
> 1st approach? Is there a standard way to handle the repeats?
>
>
> Thank you.
>
>
> Yours sincerely
> Cheng
> --
> 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