[Spam:*****] [gmx-users] REMD analysis
bharat gupta
bharat.85.monu at gmail.com
Thu May 16 17:25:13 CEST 2013
Sorry to ask this simple question but how to read the replica_index and
replica_temp files. I tried to search a lot but didn't find any
information. As I have concatenated all log files and demuxed them. Here's
first 10 lines from both files:-
replica_index:
0 0 1 2 3 4 5 6 7 8 9 10 11 12
2 1 0 2 3 4 5 6 7 9 8 10 11 12
4 1 0 2 3 4 5 6 9 7 8 10 11 12
6 1 0 2 3 4 5 9 6 7 8 10 11 12
8 1 2 0 3 4 9 5 7 6 8 10 11 12
10 1 2 3 0 9 4 7 5 6 8 11 10 12
replica_temp
0 0 1 2 3 4 5 6 7 8 9 10 11 12
2 1 0 2 3 4 5 6 7 9 8 10 11 12
4 1 0 2 3 4 5 6 8 9 7 10 11 12
6 1 0 2 3 4 5 7 8 9 6 10 11 12
8 2 0 1 3 4 6 8 7 9 5 10 11 12
10 3 0 1 2 5 7 8 6 9 4 11 10 12
On Thu, May 16, 2013 at 11:24 PM, Mark Abraham <mark.j.abraham at gmail.com>wrote:
> On Thu, May 16, 2013 at 2:04 PM, bharat gupta <bharat.85.monu at gmail.com
> >wrote:
>
> > Dear Sir,
> >
> > Here's the result of three different runs :
> >
> > Temperature distribution for three trials
> >
> > Repeat-1 280 298 317 337 359 382 406 432 460 489 520 554 589 627
> > Repeat-2 280 299 319 340 363 388 414 441 471 503 536 572 611
> > Repeat-3 280 300 322 345 370 397 426 457 490 526 564 605 649
> >
> > md.log files output from three different trials:
> >
> > Repeat-1 .37 .28 .26 .30 .25 .29 .32 .35 .32 .35 .36 .32 .31
> > Repeat-2 .30 .33 .30 .25 .19 .27 .30 .31 .27 .40 .34 .31
> > Repeat-3 .18 .22 .26 .34 .26 .28 .25 .27 .27 .25 .27 .22
> >
> > I think as the required acceptance value all the three trials are fine,
> but
> > trail 3 would be much better to continue the further runs and anlysis ??
> >
>
>
> Probably. But exchange acceptance is a poor proxy for sampling efficiency -
> see recent discussions of REMD on this list.
>
> Mark
>
> So, is it fine to continue with the third simulation ?? But still the
> > problem is that I am not getting the exact graphs with xmgrace??
> >
> >
> > On Thu, May 16, 2013 at 5:36 PM, XAvier Periole <x.periole at rug.nl>
> wrote:
> >
> > >
> > > You have to convince yourself, not me :)) But I can give you my
> opinion …
> > >
> > > On May 16, 2013, at 10:33 AM, bharat gupta <bharat.85.monu at gmail.com>
> > > wrote:
> > >
> > > > 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
> > > >>
> > > > --
> > > > 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