[gmx-users] Different results from identical tpr after MD

[Justin A. Lemkul]

Acoot Brett wrote:
> Here I want to talk on the time length of production MD.
>  
> I have a professor on computational biology. She said for MD only 
> several ps is enough. I have the expereince to run production MD for 
> several ns, and I have got stable conformation.
>  

Several ps is insufficient for all but the fastest (and thus smallest) motions 
in a macromolecule like a protein.

> But I am sure for some protein system, even a 50 ns MS will not give the 
> stable conformation, thus maybe the results are not reproducible. I am 
> not sure whether my explaination is suitable to you.
>  

The time frame necessary depends on what you're hoping to see.  Fast motions 
(like side chain rotation, etc) require short simulations.  Domain motions or 
folding simulations may require hundreds of ns or even more.  Some phenomena 
cannot be modeled efficiently by atomistic MD simulations, though other 
mesoscale approaches may perform better.

-Justin

> Cheers,
>  
> Acoot
> 
> *From:* Matthew Zwier <mczwier at gmail.com>
> *To:* Discussion list for GROMACS users <gmx-users at gromacs.org>
> *Sent:* Friday, 6 April 2012 4:24 AM
> *Subject:* Re: [gmx-users] Different results from identical tpr after MD
> 
> That's an interesting philosophical question.
> 
> In this case, you'll wind up with a 50 ns trajectory where each
> configuration is consistent with the thermodynamic ensemble you're
> approximating.  That's as close a definition to "realistic" as I think
> is worth worrying about.  You'd only need to worry about mdrun not
> giving you exactly reproducible trajectories to 7 or 8 significant
> figures of precision in atomic coordinates if, simultaneously, nature
> behaved completely deterministically, you could observe a molecular
> system with that degree of precision, and your forcefield was exact;
> none of those three things are true.
> 
> That said, 50 ns seems like a very short time to simulate a complex
> system like protein/DNA/RNA.  I'd be very worried about drawing
> anything other than observations from such a small data set (because
> you're almost certainly not converged in your sampling, as you allude
> to in your final question).
> 
> MZ
> 
> On Thu, Apr 5, 2012 at 2:08 PM, bipin singh <bipinelmat at gmail.com 
> <mailto:bipinelmat at gmail.com>> wrote:
>  > Thanks for the reply.
>  >  I read the link. So, how one can predict something reliable using these
>  > results(based on 50ns in my case) which changes on different machines?
>  > which depends more on the environment of the computer architecture 
> and other
>  > variables of mdrun rather than system (Protein/DNA/RNA) itself for a 
> short
>  > simulation, where we don't have enough resources to run infinitely long
>  > simulation.
>  > And also how to believe the statement made by several research papers 
> based
>  > on unconverged simulations.
>  >
>  > On Thu, Apr 5, 2012 at 23:12, Justin A. Lemkul <jalemkul at vt.edu 
> <mailto:jalemkul at vt.edu>> wrote:
>  >>
>  >>
>  >>
>  >> bipin singh wrote:
>  >>>
>  >>> Hi all,
>  >>>
>  >>> I am really surprised to see different results from two identical md
>  >>> simulation. I have used identical tpr files for the mdrun (for 
> 50ns) and
>  >>> after the completion of the md job I found that the results from the
>  >>> identical runs is totally different.
>  >>>
>  >>> To further confirm this, I have converted both the input tpr to mdp 
> using
>  >>> gmxdump and diff the two files and found that the mdp is identical.
>  >>>
>  >>> Please let me know what can be the reason of this behaviour. I know 
> that
>  >>> it is unexpected and even I can't believe how can it be possible.
>  >>>
>  >>
>  >> It is not unexpected at all.  Please consult:
>  >>
>  >> http://www.gromacs.org/Documentation/Terminology/Reproducibility
>  >>
>  >> -Justin
>  >>
>  >> --
>  >> ========================================
>  >>
>  >> 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 at gromacs.org 
> <mailto:gmx-users at gromacs.org>
>  >> http://lists.gromacs.org/mailman/listinfo/gmx-users
>  >> Please search the archive at
>  >> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>  >> Please don't post (un)subscribe requests to the list. Use the www
>  >> interface or send it to gmx-users-request at gromacs.org 
> <mailto:gmx-users-request at gromacs.org>.
>  >> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>  >
>  >
>  >
>  >
>  > --
>  > -----------------------
>  > Regards,
>  > Bipin Singh
>  >
>  >
>  > --
>  > gmx-users mailing list    gmx-users at gromacs.org 
> <mailto:gmx-users at gromacs.org>
>  > http://lists.gromacs.org/mailman/listinfo/gmx-users
>  > Please search the archive at
>  > http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>  > Please don't post (un)subscribe requests to the list. Use the
>  > www interface or send it to gmx-users-request at gromacs.org 
> <mailto:gmx-users-request at gromacs.org>.
>  > Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> --
> gmx-users mailing list    gmx-users at gromacs.org 
> <mailto:gmx-users at gromacs.org>
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at 
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org 
> <mailto:gmx-users-request at gromacs.org>.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> 
> 

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

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

========================================