[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