You have to convince yourself, not me :)) But I can give you my opinion …
On May 16, 2013, at 10:33 AM, bharat gupta <[email protected]> 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 <[email protected]> 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 <[email protected]> >> 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 <[email protected]> >> 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 <[email protected]> >>>> 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 <[email protected]> >>>> 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 <[email protected]> >>>>>> 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 [email protected] >>>>>>> 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 [email protected]. >>>>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >>>>>> >>>>>> -- >>>>>> gmx-users mailing list [email protected] >>>>>> 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 [email protected]. >>>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >>>>>> >>>>> -- >>>>> gmx-users mailing list [email protected] >>>>> 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 [email protected]. >>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >>>> >>>> -- >>>> gmx-users mailing list [email protected] >>>> 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 [email protected]. >>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >>>> >>> -- >>> gmx-users mailing list [email protected] >>> 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 [email protected]. >>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >> >> -- >> gmx-users mailing list [email protected] >> 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 [email protected]. >> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >> > -- > gmx-users mailing list [email protected] > 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 [email protected]. > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists -- gmx-users mailing list [email protected] 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 [email protected]. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
