[gmx-users] Gromacs 3.3.1. and 3.3.3

[Tsjerk Wassenaar]

Hi,

Fully agree with that one. But the big problem is with simulations for
comparison that you have to justify that a bug fix could not be the
cause of a difference between simulations (or could even counteract on
something which should give a difference). So even with minor
revisions, unless you can assure that the bug fix has no consequences
for the properties measured, one should not just switch within a
project. The state of the program/force field/parameters should be
part of the controlled environment, and any bias due to that can be
assumed to equally affect all simulations.
Anyway, in either case, the answer is: rerun that one simulation with
3.3.3 if you want to be able to do some meaningful comparisons (and
better repeat every simulation at least once more to improve
statistics).

Cheers,

Tsjerk


On Thu, Oct 9, 2008 at 10:16 AM, Berk Hess <gmx3 at hotmail.com> wrote:
> Hi,
>
> I agree with Tsjerk, but his mail might have not answered your question
> exaclty.
>
> In general one should not switch between major revisions in one project.
> But there should be nearly no differences in simulation results between
> minor revisions. If there are differences this means that a bug was fixed
> and you should be using the newer revision anyhow.
> So I would suggest switching to 3.3.3.
>
> Berk
>
>
> ________________________________
>> Date: Thu, 9 Oct 2008 09:27:28 +0200
>> From: tsjerkw at gmail.com
>> To: minnale_gnos at rediffmail.com; gmx-users at gromacs.org
>> Subject: Re: [gmx-users] Gromacs 3.3.1. and 3.3.3
>> CC:
>>
>> Minnale,
>>
>> Hmm, I thought I recalled answering a similar question just recently.
>> But mind you that doing statistics on MD, unless the results are very
>> obvious, you have to be as careful and consistent as can be. That
>> means that you should avoid trying to compare things which are
>> performed under different conditions, and program versions is one of
>> them. In fact, I believe there are some differences between 3.3.1 and
>> 3.3.3 which can be expected to affect the results, so in a comparison,
>> you wouldn't have any means of assessing the source of differences
>> observed. Besides that, having only single simulations makes most
>> comparisons tricky anyway. You should really try to rerun the
>> simulations to obtain an estimate of the natural range of observables
>> you can expect. In that case, you could either try randomizing program
>> versions or stick with one.
>> To cut it short, don't try to make any comparisons beyond the obvious:
>> you don't have enough statistics and you can probably not exclude an
>> effect of running with different versions of gromacs.
>>
>> As a side note, when performing simulations to compare effects of
>> simulation parameters, mutations, whatever, it is more important to
>> design and perform a set of simulations in a consistent manner than to
>> have the parameters, force field etc. perfectly apt for reproducing
>> reality. Of course the latter will help and will improve statements of
>> quantitative nature.
>>
>> Cheers,
>>
>> Tsjerk
>>
>>
>> On Thu, Oct 9, 2008 at 9:03 AM, minnale <minnale_gnos at rediffmail.com>
>> wrote:
>> >
>> > Hi Users,
>> > I have run protein system till 25ns in gromacs version 3.3.1 in dual
>> > processor system. Now I want to run new systems in cluster with gromacs
>> > version 3.3.3, in two cases the protein system is same but difference in
>> > few residues, If run new system in cluster compare the results with
>> > already
>> > ran protein system, will it make any difference?
>> >
>> >
>> > _______________________________________________
>> > gmx-users mailing list gmx-users at gromacs.org
>> > http://www.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-request at gromacs.org.
>> > Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>> >
>>
>>
>>
>> --
>> Tsjerk A. Wassenaar, Ph.D.
>> Junior UD (post-doc)
>> Biomolecular NMR, Bijvoet Center
>> Utrecht University
>> Padualaan 8
>> 3584 CH Utrecht
>> The Netherlands
>> P: +31-30-2539931
>> F: +31-30-2537623
>> _______________________________________________
>> gmx-users mailing list gmx-users at gromacs.org
>> http://www.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-request at gromacs.org.
>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>
> ________________________________
> Express yourself instantly with MSN Messenger! MSN Messenger
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.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-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>



-- 
Tsjerk A. Wassenaar, Ph.D.
Junior UD (post-doc)
Biomolecular NMR, Bijvoet Center
Utrecht University
Padualaan 8
3584 CH Utrecht
The Netherlands
P: +31-30-2539931
F: +31-30-2537623