[Spam:*****] [gmx-users] REMD analysis
bharat gupta
bharat.85.monu at gmail.com
Thu May 16 10:33:33 CEST 2013
Okay Sir, I will try two-three combinations this time and will report back
to you ...
On Thu, May 16, 2013 at 5:25 PM, XAvier Periole <x.periole at rug.nl> wrote:
>
> An acceptance ratio of 0.2/0.3 is normally best. The problem with high
> acceptance ratio is that it means that a large portion of the exchanges are
> just back and forth exchanges between consecutive exchange and are thus
> disturbing the system more that actually helping sampling.
>
> I do not know particularly the paper you mention but if you like what they
> do, it is your choice at the end.
>
> Why don;t you just increase the spacing between the replicas? You will
> need less replicas and potentially you could run two simulations instead of
> one and evaluate the convergence ...
>
> On May 16, 2013, at 1:50 AM, bharat gupta <bharat.85.monu at gmail.com>
> wrote:
>
> > The plots that I showed in my last mail were for all replicas. I tried
> > plotting the first 500 ps of replica_index and replica_time files. I
> think
> > the plots look fine, and there could be problem with the plotting tool .
> > Here the link for both files ,
> > https://www.dropbox.com/s/2g16mlxfsme4rx2/replica_temp.bmp
> > https://www.dropbox.com/s/8jfs0b9whu6j7lo/replica_index.bmp
> >
> > Now regarding the high acceptance ratio which is 0.5 , I came across a
> > paper (http://www.pnas.org/content/100/13/7587.full.pdf), here they have
> > mentioned that their average acceptance ratio ranged between 30 to 80%. I
> > have a question here, how did they calculate the range for the average
> > acceptance ratio or is it average ratio for each replica . Actually, this
> > is the reference I am following. I am also interested in peptide folding
> > simulation, similar to this article.
> >
> > I want to know, whether the average acceptance ratio that I have got for
> my
> > trial simulation is correct , together with the replica_temp and
> > replica_remd plots. Can I proceed for large production runs to complete
> my
> > experiment ??
> >
> >
> >
> >
> > On Tue, May 14, 2013 at 6:34 PM, XAvier Periole <x.periole at rug.nl>
> wrote:
> >
> >>
> >> The interval between the exchange trial affect the efficiency of REMD
> but
> >> not the the exchange ratio (at least in principle).
> >>
> >> In you case I am not sure what the plot are showing! Are these showing
> all
> >> the replicas? what are the units?
> >>
> >> On May 14, 2013, at 5:07 AM, bharat gupta <bharat.85.monu at gmail.com>
> >> wrote:
> >>
> >>> Dear Sir,
> >>>
> >>> Here's the result for the REMD trial with large temperature gaps.
> >>>
> >>> Temp. distribution : 280.0 294.9 310.7 327.3 344.7 363.1 382.5 402.9
> >> 424.4
> >>> 447.1 471.0 496.1 522.6 550.5 579.9 610.8
> >>>
> >>> Out of md16.log :
> >>>
> >>> Replica exchange statistics
> >>> Repl 249 attempts, 125 odd, 124 even
> >>> Repl average probabilities:
> >>> Repl 0 1 2 3 4 5 6 7 8 9 10 11 12
> >>> 13 14 15
> >>> Repl .40 .34 .38 .43 .43 .36 .45 .40 .37 .48 .47 .45
> .47
> >>> .44 .46
> >>>
> >>> Repl number of exchanges:
> >>> Repl 0 1 2 3 4 5 6 7 8 9 10 11 12
> >>> 13 14 15
> >>> Repl 50 42 46 52 57 40 58 49 42 53 61 63
> 56
> >>> 57 58
> >>>
> >>> Repl average number of exchanges:
> >>> Repl 0 1 2 3 4 5 6 7 8 9 10 11 12
> >>> 13 14 15
> >>> Repl .40 .34 .37 .42 .46 .32 .46 .40 .34 .43 .49 .51
> .45
> >>> .46 .46
> >>> Average acceptance ratio : 0.46
> >>>
> >>> But, the repli_index.xvg and replica_temp.xvg files still shows that
> the
> >>> replicas does not exchange equally well .
> >>>
> >>> https://www.dropbox.com/s/zkbwpuj7l2o282b/replica_index.png
> >>> https://www.dropbox.com/s/0c8gp584v1hvlbx/replica_temp.png
> >>>
> >>> what could be wrong in this case?? Is it the mdp file settings or
> >> implicit
> >>> solvent setting. Does the time to replica to exhange also affects their
> >>> swapping ??
> >>>
> >>>
> >>>
> >>> On Tue, May 14, 2013 at 12:24 AM, XAvier Periole <x.periole at rug.nl>
> >> wrote:
> >>>
> >>>>
> >>>> You need to increase the temperature gaps indeed if you want
> acceptance
> >>>> ratio ~0.2/0.3. But again this won't work with the water …
> >>>>
> >>>> It is not clear what happens in your index file but probably a problem
> >>>> from grace to plot so many points … you can try to increase the "Max
> >>>> drawing path length" in the preference menu of grace.
> >>>>
> >>>> On May 13, 2013, at 4:22 PM, bharat gupta <bharat.85.monu at gmail.com>
> >>>> wrote:
> >>>>
> >>>>> Dear Sir,
> >>>>>
> >>>>> I repeated the simulation again for 25 replicas with the following
> >> temp.
> >>>>> distribution .
> >>>>>
> >>>>> 280
> >>>>> 289.1
> >>>>> 298.5
> >>>>> 308.2
> >>>>> 318.2
> >>>>> 328.6
> >>>>> 339.3
> >>>>> 350.3
> >>>>> 361.7
> >>>>> 373.5
> >>>>> 385.6
> >>>>> 398.1
> >>>>> 411.1
> >>>>> 424.4
> >>>>> 438.3
> >>>>> 452.5
> >>>>> 467.2
> >>>>> 482.4
> >>>>> 498.1
> >>>>> 514.3
> >>>>> 531.0
> >>>>> 548.3
> >>>>> 566.1
> >>>>> 584.5
> >>>>> 603.5
> >>>>> 623.2
> >>>>>
> >>>>> The output of md.log file is :-
> >>>>>
> >>>>> Replica exchange statistics
> >>>>> Repl 24999 attempts, 12500 odd, 12499 even
> >>>>> Repl average probabilities:
> >>>>> Repl 0 1 2 3 4 5 6 7 8 9 10 11
> 12
> >>>>> 13 14 15 16 17 18 19 20 21 22 23 24 25
> >>>>> Repl .63 .63 .62 .62 .61 .61 .60 .60 .59 .59 .58 .59
> >> .59
> >>>>> .60 .60 .61 .62 .62 .63 .64 .64 .65 .65 .66 .66
> >>>>>
> >>>>> Repl number of exchanges:
> >>>>> Repl 0 1 2 3 4 5 6 7 8 9 10 11
> 12
> >>>>> 13 14 15 16 17 18 19 20 21 22 23 24 25
> >>>>> Repl 7822 7752 7816 7760 7639 7628 7511 7442 7375 7332 7312 7424
> >> 7408
> >>>>> 7410 7522 7559 7684 7697 7878 7927 7917 8073 8151 8208 8266
> >>>>>
> >>>>> Repl average number of exchanges:
> >>>>> Repl 0 1 2 3 4 5 6 7 8 9 10 11
> 12
> >>>>> 13 14 15 16 17 18 19 20 21 22 23 24 25
> >>>>> Repl .63 .62 .63 .62 .61 .61 .60 .60 .59 .59 .58 .59
> >> .59
> >>>>> .59 .60 .60 .61 .62 .63 .63 .63 .65 .65 .66 .66
> >>>>>
> >>>>> The average acceptance ration is around 0.6 which is still high.
> >>>>>
> >>>>> The link for replica_temp,replica_index :
> >>>>> https://www.dropbox.com/s/c7soajnwc3uww8j/replica_temp.png
> >>>>> https://www.dropbox.com/s/wvx82m4c6cnsfit/replica_index.png
> >>>>>
> >>>>> The temp files look better but the index file looks weird ...
> >>>>>
> >>>>> Do i need to experiment with the gap difference in order to get the
> >>>>> required ration of 0.2-0.3 ?? There is some problem with the .mdp
> file
> >>>>> settings??
> >>>>>
> >>>>> --
> >>>>> Bharat
> >>>>> --
> >>>>> gmx-users mailing list 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.
> >>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >>>>
> >>>> --
> >>>> gmx-users mailing list 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.
> >>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >>>>
> >>> --
> >>> gmx-users mailing list 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.
> >>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >>
> >> --
> >> gmx-users mailing list 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.
> >> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >>
> > --
> > gmx-users mailing list 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.
> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> --
> gmx-users mailing list 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.
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
More information about the gromacs.org_gmx-users
mailing list