[gmx-users] Replica exchange probabilities beyond version 2016
Pritam Ganguly
pganguly at chem.ucsb.edu
Wed Apr 15 02:37:06 CEST 2020
Hi Gül,
You are absolutely correct! I just checked this with tau_t=10, and it
brought back the old behavior! Thank you so much for your suggestion!
Probabilities with tau_t=10, NH thermostat:
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 26 27
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
58 59 60 61 62 63
Repl .29 .29 .30 .29 .30 .28 .31 .30 .29 .28 .29 .26
.28 .29 .28 .27 .29 .26 .27 .28 .26 .26 .28 .27 .26 .27
.26 .27 .26 .25 .26 .26 .23 .27 .26 .23 .24 .27 .23 .24
.28 .27 .26 .25 .27 .24 .28 .24 .27 .27 .27 .27 .28 .25
.27 .27 .26 .28 .26 .28 .27 .28 .30
Thanks again!
Regards,
Pritam
On 4/13/20 7:55 PM, Gül Zerze wrote:
> Hi Pritam,
>
> Did you check your energy fluctuations? Looks like you are using NH
> thermostat, you might be suffering from the interaction of NH
> thermostat oscillations with resonance modes in the system, which
> makes energy fluctuations substantially larger (compared to
> fluctuations with NH in older versions of Gromacs). As a result of
> amplified fluctuations, you have higher exchange probability. See here
> for a previous relevant exchange:
> https://redmine.gromacs.org/issues/2904 For my system, making tau_t=10
> helped to recover the old behavior.
>
> Best,
> Gül
> *--*
> *Gül H. Zerze*
> Postdoctoral Research Associate
>
> Chemical and Biological Engineering
> Princeton University
> http://pablonet.princeton.edu/
>
>
>
> On Mon, Apr 13, 2020 at 8:01 PM Pritam Ganguly <pganguly at chem.ucsb.edu
> <mailto:pganguly at chem.ucsb.edu>> wrote:
>
> Hello,
>
> I have noticed that the average exchange probability for a replica
> exchange run increases significantly if I use Gromacs versions
> 2018, as
> compared to versions 2016 or earlier. I have tested that by
> simulating
> the same system with versions 2016.4 and 2018.3 and the average
> probabilities I find are:
>
> Version 2016.4:
>
> Replica exchange statistics
> Repl 1666 attempts, 833 odd, 833 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 26 27
> 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
> 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
> 58 59 60 61 62 63
> Repl .29 .33 .29 .30 .28 .30 .30 .29 .29 .29 .30 .26
> .29 .30 .28 .29 .26 .26 .29 .28 .26 .27 .27 .27 .27 .28
> .26 .27 .26 .25 .27 .25 .26 .27 .25 .24 .25 .26 .27 .25
> .27 .26 .26 .26 .25 .26 .26 .24 .25 .27 .26 .27 .26 .25
> .27 .27 .28 .26 .27 .26 .27 .27 .28
>
>
> Version 2018.3:
>
> Replica exchange statistics
> Repl 1666 attempts, 833 odd, 833 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 26 27
> 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
> 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
> 58 59 60 61 62 63
> Repl .39 .39 .37 .42 .38 .38 .39 .36 .36 .38 .40 .39
> .40 .39 .37 .35 .40 .39 .35 .38 .40 .35 .38 .35 .36 .35
> .36 .36 .38 .36 .37 .36 .35 .36 .36 .37 .35 .35 .35 .37
> .37 .35 .37 .36 .37 .34 .36 .38 .38 .38 .35 .37 .36 .39
> .36 .40 .38 .38 .40 .36 .37 .36 .33
>
> These are test runs for 5 ns with exchange frequencies of 3 ps. I
> have
> used the same input parameters for both of these runs (mdp is
> attached).
> I actually observed this with many other systems and I was looking
> for
> any relevant gromacs release note for versions 2018 or beyond
> related to
> the replica exchange algorithm or the way the energies are
> calculated,
> but I could not figure it out.
>
> Regards,
>
> Pritam
>
> --
> Pritam Ganguly
> Department of Chemistry and Biochemistry
> University of California at Santa Barbara
> Santa Barbara, CA, USA 93106
>
> Office: PSBN 3606
> Phone: +1-805-893-2767
> http://people.chem.ucsb.edu/ganguly/pritam
>
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users
> or send a mail to gmx-users-request at gromacs.org
> <mailto:gmx-users-request at gromacs.org>.
>
--
Pritam Ganguly
Department of Chemistry and Biochemistry
University of California at Santa Barbara
Santa Barbara, CA, USA 93106
Office: PSBN 3606
Phone: +1-805-893-2767
http://people.chem.ucsb.edu/ganguly/pritam
More information about the gromacs.org_gmx-users
mailing list