[gmx-users] Hamiltonian replica exchange with solute tempering in 5.x version (bug?)

[Christopher Neale]

I don't think it is a bug per se, more like something the developers didn't consider. The HREX fork of the plumed code (which only works right now for v. 4 of gromacs) does what you want (and it does it a lot faster if your heated group is large since the gromacs free energy code is inexplicable slow when decoupling large groups). I'll note that the users list for this plumed code does mention the same issue that you are experiencing, namely that gromacs itself thinks there is nothing different between the replicas and therefore fails -- which is probably the main reasons that this REST fork of plumed was created in the first place.

Chris.

________________________________________
From: gromacs.org_gmx-users-bounces at maillist.sys.kth.se <gromacs.org_gmx-users-bounces at maillist.sys.kth.se> on behalf of 呂宗諭 <joyeuxnoel8 at gmail.com>
Sent: 13 May 2015 11:29
To: gromacs.org_gmx-users at maillist.sys.kth.se
Subject: [gmx-users] Hamiltonian replica exchange with solute tempering in      5.x version (bug?)

Dear all,

I would appreciate you to give me any comments or share your experience on
the question below.

I am trying to run Hamiltonian replica exchange with solute tempering
(H-REST) in either 5.02 or 5.04 version to speed up the conformation
sampling of my protein along with a caging small molecule. I made some
changes in the force field parameters and used free energy function to
effectively realize H-REST.
However, after I prepared my equilibrated systems with different lambda
values and tried to run H-REST, mdrun keeps saying that properties of every
system are all the same.
I also found a weird thing, not sure if it's a bug. All of the mdout.mdp
files from different equilibrated systems have different lambda values at
the beginning. However, after they are combined to run H-REST, the .log
files turn out to have identical lambda values. The values are always that
of the last replica (higher effective temperature).

Here are an example of the mdp options I tried and the outcomes (.log file).
Any other variants will still just make every replica the same as the last
replica.

mdp option for replica 1

Free-energy = yes

Init-lambda-state = 0

fep-lambdas = 0

coul-lambdas = *0*

couple-moltype = Other2

nstdhdl = 0

couple-lambda0 = vdw-q

couple-lambda1 = vdw

couple-intramol = yes

mdp option for replica 8

Free-energy = yes
Init-lambda-state = 0
fep-lambdas = 0
coul-lambdas = *0.577*
couple-moltype = Other2
nstdhdl = 0
couple-lambda0 = vdw-q
couple-lambda1 = vdw
couple-intramol = yes


.log file of replica 1

        fep-lambdas =            0

       mass-lambdas =            0

       coul-lambdas =        *0.577*

        vdw-lambdas =            0

     bonded-lambdas =
  restraint-lambdas =            0

temperature-lambdas =            0


.log file of replica 8

        fep-lambdas =            0

       mass-lambdas =            0

       coul-lambdas =        *0.577*

        vdw-lambdas =            0

     bonded-lambdas =            0

  restraint-lambdas =            0

temperature-lambdas =            0

I've also tried the following mdp option according to Michael's advice.
https://mailman-1.sys.kth.se/pipermail/gromacs.org_gmx-users/2013-August/083430.html

Free-energy = yes
Init-lambda-state = 0
fep-lambdas = 0 0 0 0 0 0 0 0
coul-lambdas = 0 0.163 0.280 0.369 0.438 0.493 0.539 0.577
dhdl-derivatives = no
couple-moltype = Other2
nstdhdl = 0
couple-lambda0 = vdw-q
couple-lambda1 = vdw
couple-intramol = yes

but still not working.

If this is really a bug not just a mistake of my wrong settings, maybe I
should try PLUMED plug-in to invoke H-REST instead.

Thanks for your reading
Best regards,

Tony
--
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.