[gmx-users] REMD analysis
XAvier Periole
x.periole at rug.nl
Thu May 16 16:10:04 CEST 2013
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
More information about the gromacs.org_gmx-users
mailing list