[Spam:*****] [gmx-users] REMD analysis
XAvier Periole
x.periole at rug.nl
Thu May 16 10:36:01 CEST 2013
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
More information about the gromacs.org_gmx-users
mailing list