[gmx-users] REMD analysis

bharat gupta bharat.85.monu at gmail.com
Thu May 16 16:15:57 CEST 2013 

Okay, now I can start with large production runs .


On Thu, May 16, 2013 at 11:10 PM, XAvier Periole <x.periole at rug.nl> wrote:

>
> Indeed the Repeat-3 seems good. But I would guess you did not run too
> long, right! That would explain the distribution of values!
>
> On 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 ??
> >
> > 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