questions about HREX tutorial

[Shyno Mathew]

Dear all,

I have few questions regarding the tests mentioned in the HREX tutorial:

https://github.com/GiovanniBussi/plumed2/blob/v2.1-hrex/user-doc/tutorials/hrex.txt

1. I did a 1 ns run of my system, then did a rerun on this trajectory with the scaled topology file, lambda=1

    I understand all the energies has to be equal. However, I see a difference in proper dihedrals.

    Please see page 1 of attached file, the plot on left is using the raw data.

    The plot on right shows the difference between proper dihedral energy values at lambda=1 and unscaled

    It varies between 2.5 and 22.5, can I ignore this or is it a serious issue?

2. When I compare lambda=1 to lambda= 0.5, both are reruns of an existing trajectory (please note this is not the trajectory used for question 1, this was done earlier), the proper dihedral energy values at lambda=1 is twice the values at lambda=0.5, which should be the case.

In this case, the difference is in the coulomb values, please see pages 2,3

The difference "(values at lambda=1)- (2* values at lambda=0.5)" varies between 0.045 and 0.075 (for short range Coulomb, 0.056 and 0.058 (for Coulomb-14),  -0.05 and -0.02 (for Recip. coulomb)

Again is it ok to ignore these?

3. I have two versions of gromacs installed: gromacs 4.5.5, gromacs 4.6.7 (installed for plumed)

    When I use "grompp_d" or "mdrun_d" I am using the ones from gromacs 4.5.5

    Is that why I am noticing difference in the energy values?

Any help will be really appreciated

thanks,

S

[Shyno Mathew]

Hi everyone,

To avoid the issue of two versions of gromacs, I re-installed gromacs-4.6.7 and added the plumed patch.

However I can't see "plumed partial_tempering" any more, all I have is  "plumed-partial_tempering"
When I issue the following command:
plumed partial_tempering 1.0 < 02processed.top > topol-scaled.top

I get the message: "NOTE: shell only version, useful when plumed is cross compiled"

If I compare the  topology files. the one being made with current plumed vs. the previous one, the epsilon 1-4 values are being removed!
For example:

values from the topology file using previous installation:

[ pairtypes ]
; i     j       func    sigma1-4        epsilon1-4 ; THESE ARE 1-4 INTERACTIONS
CP1     CP1     1       0.338541512893  0.04184
CP1_    CP1_    1       0.338541512893  0.041840000000    ; scaled


Now it says scaled, but the epsilon 1-4 values are being removed!
[ pairtypes ]
; i     j       func    sigma1-4        epsilon1-4 ; THESE ARE 1-4 INTERACTIONS
CP1     CP1     1       0.338541512893  0.04184
CP1_    CP1_    1       0.338541512893  0.000000000000    ; scaled


Can someone help please?

thanks,
Shyno

[Giovanni Bussi]

Hi

On Fri, May 22, 2015 at 6:17 PM, Shyno Mathew <shyno...@gmail.com> wrote:

Dear all,

I have few questions regarding the tests mentioned in the HREX tutorial:

https://github.com/GiovanniBussi/plumed2/blob/v2.1-hrex/user-doc/tutorials/hrex.txt

1. I did a 1 ns run of my system, then did a rerun on this trajectory with the scaled topology file, lambda=1

    I understand all the energies has to be equal. However, I see a difference in proper dihedrals.

    Please see page 1 of attached file, the plot on left is using the raw data.

    The plot on right shows the difference between proper dihedral energy values at lambda=1 and unscaled

    It varies between 2.5 and 22.5, can I ignore this or is it a serious issue?

Difference seems small with respect to total value, but large as absolute numbers. Anyway, notice that if you do a rerun from an xtc positions will not have full precision. Better to compare two reruns made with two different tpr files as you did later. Or you should fuse full precision trajectories.

 

2. When I compare lambda=1 to lambda= 0.5, both are reruns of an existing trajectory (please note this is not the trajectory used for question 1, this was done earlier), the proper dihedral energy values at lambda=1 is twice the values at lambda=0.5, which should be the case.

In this case, the difference is in the coulomb values, please see pages 2,3

The difference "(values at lambda=1)- (2* values at lambda=0.5)" varies between 0.045 and 0.075 (for short range Coulomb, 0.056 and 0.058 (for Coulomb-14),  -0.05 and -0.02 (for Recip. coulomb)

Again is it ok to ignore these?

I guess you can ignore these difference. They could come from finite precision in top files, and anyway are << kbT for the total energy, which is fine.

 

3. I have two versions of gromacs installed: gromacs 4.5.5, gromacs 4.6.7 (installed for plumed)

    When I use "grompp_d" or "mdrun_d" I am using the ones from gromacs 4.5.5

    Is that why I am noticing difference in the energy values?

It could be that different gromacs version give you slightly different energies. So better to compare energies computed with the same version.

Giovanni 

Any help will be really appreciated

thanks,

S

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/67d2c5df-ee2f-4835-8e8e-4d58e21a91ab%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Giovanni Bussi]

On Sun, May 24, 2015 at 5:55 PM, Shyno Mathew <shyno...@gmail.com> wrote:

Hi everyone,

To avoid the issue of two versions of gromacs, I re-installed gromacs-4.6.7 and added the plumed patch.

However I can't see "plumed partial_tempering" any more, all I have is  "plumed-partial_tempering"

This is weird. Did you reinstall plumed? Did you reinstall the proper plumed version? Notice that standard plumed does not contain the partial_tempering script. If you are familiar with git you might want to merge v2.1-hrex with v2.1 from official repo to have latest patches. Anyway this should not be a problem here. Just check if "plumed" executable is the version from hrex branch (which should contain the partial_tempering command) or the one from the official repo.

 

When I issue the following command:
plumed partial_tempering 1.0 < 02processed.top > topol-scaled.top

I get the message: "NOTE: shell only version, useful when plumed is cross compiled"

Many of the "plumed xxx" commands are also provided as "plumed-xxx" so as to allow their usage in cross-compiled environment. The two versions should provide identical results. But it is anyway strange that only one is available on your system.

 

If I compare the  topology files. the one being made with current plumed vs. the previous one, the epsilon 1-4 values are being removed!
For example:

values from the topology file using previous installation:

[ pairtypes ]
; i     j       func    sigma1-4        epsilon1-4 ; THESE ARE 1-4 INTERACTIONS
CP1     CP1     1       0.338541512893  0.04184
CP1_    CP1_    1       0.338541512893  0.041840000000    ; scaled


Now it says scaled, but the epsilon 1-4 values are being removed!
[ pairtypes ]
; i     j       func    sigma1-4        epsilon1-4 ; THESE ARE 1-4 INTERACTIONS
CP1     CP1     1       0.338541512893  0.04184
CP1_    CP1_    1       0.338541512893  0.000000000000    ; scaled


Can someone help please?

This looks weird. The awk script seems not to work. Either you are setting the scaling factor to 0 (i.e. you did not pass it properly on the command line), or perhaps awk is not interpreting numbers correctly because of some environment variable. Perhaps your computer is not set to use English as language and expects real number to use commas?

Giovanni

 

thanks,
Shyno

On Friday, May 22, 2015 at 12:17:37 PM UTC-4, Shyno Mathew wrote:

Dear all,

I have few questions regarding the tests mentioned in the HREX tutorial:

https://github.com/GiovanniBussi/plumed2/blob/v2.1-hrex/user-doc/tutorials/hrex.txt

1. I did a 1 ns run of my system, then did a rerun on this trajectory with the scaled topology file, lambda=1

    I understand all the energies has to be equal. However, I see a difference in proper dihedrals.

    Please see page 1 of attached file, the plot on left is using the raw data.

    The plot on right shows the difference between proper dihedral energy values at lambda=1 and unscaled

    It varies between 2.5 and 22.5, can I ignore this or is it a serious issue?

2. When I compare lambda=1 to lambda= 0.5, both are reruns of an existing trajectory (please note this is not the trajectory used for question 1, this was done earlier), the proper dihedral energy values at lambda=1 is twice the values at lambda=0.5, which should be the case.

In this case, the difference is in the coulomb values, please see pages 2,3

The difference "(values at lambda=1)- (2* values at lambda=0.5)" varies between 0.045 and 0.075 (for short range Coulomb, 0.056 and 0.058 (for Coulomb-14),  -0.05 and -0.02 (for Recip. coulomb)

Again is it ok to ignore these?

3. I have two versions of gromacs installed: gromacs 4.5.5, gromacs 4.6.7 (installed for plumed)

    When I use "grompp_d" or "mdrun_d" I am using the ones from gromacs 4.5.5

    Is that why I am noticing difference in the energy values?

Any help will be really appreciated

thanks,

S

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/b3ca8596-b55f-4028-8f5d-0391a068eea3%40googlegroups.com.

[Shyno Mathew]

Hi,

For plumed, I used the file I downloaded on March 11 from the hrex branch
I downloaded it from "https://github.com/GiovanniBussi/plumed2/tree/v2.1-hrex"
Also, I can see partial_tempering.sh here:
/home/shyno/GromacswPLUMED/plumed2-2.1-hrex/scripts

Following are all the plumed commands I have

plumed
plumed-mklib               
plumed-newcv
plumed-partial_tempering
plumed-patch
plumed-runtime
plumed-shared


Also, I just did the regression tests:
cd regtest
make

here is the last few lines of the  output
+ Final report:
+ 95 tests performed, 25 tests not appliable
+ 0 errors found
+ Well done!!
+++++++++++++++++++++++++++++++++++++++++++++++++++++


scripts/what
WARNING : There are no regtests for action COMMITTOR
WARNING : There are no regtests for action CH3SHIFTS
WARNING : There are no regtests for action FAKE
WARNING : There are no regtests for action MULTI-RMSD
WARNING : There are no regtests for action Q3
WARNING : There are no regtests for action LOCAL_Q3
WARNING : There are no regtests for action LOCAL_Q4
WARNING : There are no regtests for action FUNCSUMHILLS
WARNING : There are no regtests for action TARGET
WARNING : There are no regtests for action RANDOM_EXCHANGES
WARNING : There are no regtests for action IMD
WARNING : There are no regtests for action TORSIONS
WARNING : There are no regtests for action XDISTANCES
WARNING : There are no regtests for action YDISTANCES
WARNING : There are no regtests for action ZDISTANCES
WARNING : There are no regtests for action GHOST


thanks,
Shyno

[Giovanni Bussi]

Problem 1 (missing plumed partial_tempering):

Check the output of

plumed info --root

In that directory you should find a dir scripts/ containing partial_tempering.sh

If so, then plumed partial_tempering should be available as a command

Problem 2 (LJ epsilon set to zero):

Are you sure you are passing the correct scaling argument to the script? notice that some of the regtests in the hrex branch actually check that partial_tempering.sh is working properly, so if you don't see failures it should work. Are you sure you are not mixing together two different plumed versions?

Inviato da iPhone

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/b6d50e6f-152f-45df-9de1-001d2456d69a%40googlegroups.com.

[Shyno Mathew]

Problem 1 (missing plumed partial_tempering):

yes, partial_tempering.sh is available in scripts directory (/home/shyno/GromacswPLUMED/plumed2-2.1-hrex/scripts)
Also, I checked  /plumed2-2.1-hrex/src/lib
the executables available here are:

plumed-mklib
plumed-patch
plumed-partial_tempering
 plumed-newcv
 plumed-runtime
plumed
libplumed.so
libplumedKernel.so
plumed-shared

Problem 2 (LJ epsilon set to zero):

You are correct, I wasn't passing the correct value to awk scripts as part of the CMAP correction, was calling it $lambda although I put the correct value here "plumed-partial_tempering 1.0 < 02processed.top | awk script"
Now my topology files are identical :)

thanks so much,
Shyno

[Shyno Mathew]

Hi all,

I am trying do REST2 with the following:

Tmin = 300 K, Tmax= 600 K

nrep=5 

I made different topology files for different lambda values.

Now for making different .tpr files, I use different .mdp files?

These .mdp files only differ in the temperature, correct?

thanks,

Shyno

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/0utHdmEvpnM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/53d7d4fe-c907-4e34-93a8-70c13cdec41a%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Sincerely
Shyno Mathew

[Shyno Mathew]

Hi all,

I am also having another issue: have two versions of gromacs installed gromacs 4.6.5 and gromacs 4.6.7
When I execute "mdrun_mpi" it's uses gromacs 4.6.5, and not gromacs 4.6.7
I had to reinstall gromacs 4.6.7 to have the MPI version, but sill this issue remains

For eg. if I do:
mpirun -np 5 mdrun_mpi -v -deffnm topol -plumed plumed.dat -multi 5 -replex 100 -hrex

I get the error:

Invalid command line argument:
-plumed

thanks,
Shyno

To unsubscribe from this group and all its topics, send an email to plumed-users+unsubscribe@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/53d7d4fe-c907-4e34-93a8-70c13cdec41a%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Sincerely
Shyno Mathew

[Giovanni Bussi]

"Invalid command line argument:
-plumed"

means that you are using a version of gromacs that was not patched with plumed

About the different tpr files for Hamiltonian replica exchange, notice that:

1. They should have the same temperature, though force fields scaled with a different lambda

2. be sure that gromacs accepts the "-hrex" flag, so as to confirm that you are using the proper version of gromacs

Giovanni

To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/53d7d4fe-c907-4e34-93a8-70c13cdec41a%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Sincerely
Shyno Mathew

--

You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/cc025d69-c452-4de5-9899-13c9b124cd0c%40googlegroups.com.

[Shyno Mathew]

Hi all,

I was able to fix the issue with mdrun_mpi, now gromacs4.6.7 is correctly intsalled and patched with plumed (MPI version)
Now I encounter some other errors.
When I execute "mpirun -np 4 mdrun_mpi -v -plumed plumed.dat -deffnm topol -multi 4 -replex 100 -hrex"
I get the error:
The number of nodes (1) is not a multiple of the number of simulations (4)

But if I do a regular run with plumed it runs fine, this means MPI is working!
For eg. if I execute, "mpirun -np 4 mdrun_mpi -v -deffnm topol0 -plumed plumed.dat" it runs fine.

Also, the plumed.dat file I am using is an empty file.

thanks,
Shyno

[Shyno Mathew]

Hello everyone,

Can someone help with the following error:

When I execute "mpirun -np 4 mdrun_mpi -v -plumed plumed.dat -deffnm topol -multi 4 -replex 100 -hrex"
I get the error:
The number of nodes (1) is not a multiple of the number of simulations (4)

But if I do a regular run with plumed it runs fine, so MPI is working!

For eg. if I execute, "mpirun -np 4 mdrun_mpi -v -deffnm topol0 -plumed plumed.dat" it runs fine.

Also, the plumed.dat file I am using is an empty file.

thanks,
Shyno

[Giovanni Bussi]

The hrex patch is more tricky than normal plumed.

Can you try with normal plumed and replica exchange first?

Giovanni

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/a4f855b7-324a-40a3-a349-f4e4f0f2a01a%40googlegroups.com.

[Shyno Mathew]

ok, i tried the following:

mpirun -np 4 mdrun_mpi -v -plumed plumed.dat -deffnm topol -multi 4 -replex 100

same error "The number of nodes (1) is not a multiple of the number of simulations (4)"

thanks,
Shyno

[Giovanni Bussi]

Did you try without plumed?

mpirun -np 4 mdrun_mpi -v -deffnm topol -multi 4 -replex 100

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/36384582-13f1-40a0-9e08-64605b52b9df%40googlegroups.com.

[Shyno Mathew]

yes, still the same error. I am trying these simulations in my local machine which has 4 processors, unfortunately our cluster is down!

I just got access to another cluster, so I am going to set up these simulations there

thanks,

Shyno

[Shyno Mathew]

Hi all,

So I am trying to patch gromacs 4.6.7 with plumed in a shared cluster. While compiling I get the message " configure: **** PLUMED will be compiled using MPI" but later when I tried to patch, I get the message: “PLUMED is compiled WITHOUT MPI support so you CANNOT configure gromacs-4.6.7 with MPI”

Please take a look at the steps I followed and the messages I am getting:

1. downloaded gromacs 4.6.7

    tar zxvf gromacs-4.6.7.tar.gz

 

2. unzip plumed2-2.1-hrex.zip

    cd plumed2-2.1-hrex

    ./configure --prefix=/vega/cheme/users/sm3334                   

 (I am doing this in a shared cluster, so don't have sudo access, or can't write/usr/local/lib)

 

message

configure: PLUMED seems to be configured properly!

configure: **************************

configure: I will now check if C++ objects can be linked by C/Fortran compilers

configure: This is relevant if you want to use plumed patch --static on a non-C++ code

checking whether C can link a C++ object... no

checking whether C can link a C++ object with library -lstdc++... yes

checking whether FORTRAN can link a C++ object... yes

configure: **** PLUMED will be installed using prefix /vega/cheme/users/sm3334

configure: **** You can change this later setting PLUMED_PREFIX environment variable before installing

configure: **** Executable will be named 'plumed'

configure: **** To add a suffix to this name, set PLUMED_LIBSUFFIX environment variable before installing

configure: **** PLUMED will be compiled using MPI

configure: creating ./config.status

config.status: creating Makefile.conf

config.status: creating sourceme.sh

 

3. make -j 4

message:
chmod a+x plumed-patch

g++  ../main/main.o ../wrapper/Plumed.o -o plumed-runtime -ldl

Building Plumed.inc, runtime-linking version

Building Plumed.inc, shared version

Building Plumed.cmake, runtime-linking version

Building Plumed.cmake, shared version

Building Plumed.inc, static version

Building Plumed.cmake, static version

g++  ../main/main.o /vega/cheme/users/sm3334/GROMACSwPLUMED/plumed2-2.1-hrex/src/lib/libplumed.so -o plumed-shared -ldl

make[4]: Leaving directory `/vega/cheme/users/sm3334/GROMACSwPLUMED/plumed2-2.1-hrex/src/lib'

make[3]: Leaving directory `/vega/cheme/users/sm3334/GROMACSwPLUMED/plumed2-2.1-hrex/src/lib'

make[2]: Leaving directory `/vega/cheme/users/sm3334/GROMACSwPLUMED/plumed2-2.1-hrex/src'

make[1]: Leaving directory `/vega/cheme/users/sm3334/GROMACSwPLUMED/plumed2-2.1-hrex'

 

4. make install

 

message

*** PLUMED has been installed ***

 

Install prefix : /vega/cheme/users/sm3334

Full name      : plumed

 

Setup your environment

- Ensure this is in your execution path : /vega/cheme/users/sm3334/bin

- Ensure this is in your include path   : /vega/cheme/users/sm3334/include

- Ensure this is in your library path   : /vega/cheme/users/sm3334/lib

For runtime binding:

- Set this environment variable         : PLUMED_KERNEL=/vega/cheme/users/sm3334/lib/libplumedKernel.so

 

To create a tcl module that sets all the variables above, use this one as a starting point:

/vega/cheme/users/sm3334/lib/plumed//src/lib/modulefile

 

To uninstall, remove the following files and directories:

/vega/cheme/users/sm3334/lib/plumed/

/vega/cheme/users/sm3334/include/plumed

/vega/cheme/users/sm3334/bin/plumed

/vega/cheme/users/sm3334/bin/plumed-mklib

/vega/cheme/users/sm3334/bin/plumed-newcv

/vega/cheme/users/sm3334/bin/plumed-partial_tempering

/vega/cheme/users/sm3334/bin/plumed-patch

/vega/cheme/users/sm3334/lib/libplumed.so

/vega/cheme/users/sm3334/lib/libplumedKernel.so

make[3]: Leaving directory `/vega/cheme/users/sm3334/GROMACSwPLUMED/plumed2-2.1-hrex/src/lib'

make[2]: Leaving directory `/vega/cheme/users/sm3334/GROMACSwPLUMED/plumed2-2.1-hrex/src/lib'

make[1]: Leaving directory `/vega/cheme/users/sm3334/GROMACSwPLUMED/plumed2-2.1-hrex/src'

 

5. cd ../gromacs-4.6.7

    source ../plumed2-2.1-hrex/sourceme.sh

    plumed patch –p --shared

message:

Linking Plumed.h and Plumed.inc (shared mode)

Patching with on-the-fly diff from stored originals

patching file ./src/kernel/CMakeLists.txt

patching file ./src/kernel/md.c

patching file ./src/kernel/mdrun.c

patching file ./src/kernel/repl_ex.c

patching file ./src/kernel/repl_ex.h

patching file ./src/mdlib/force.c

patching file ./src/mdlib/minimize.c

 

PLUMED is compiled WITHOUT MPI support so you CANNOT configure gromacs-4.6.7 with MPI

 

thanks in advnace for the help,

Shyno

[Shyno Mathew]

Dear all,

I was able to install gromacs 4.6.7 with plumed patching on the shared cluster.
Now I am trying to run a regular MD run and I am getting the following error:

mdrun_mpi: error while loading shared libraries: libplumed.so: cannot open shared object file: No such file or directory
This error is repeated 32 times as I am trying to use 2 nodes (32 cores in total)

Since it's a shared cluster, I had to intsall gromacs under my username
these are the locations:
$which mdrun_mpi
/vega/cheme/users/sm3334/GROMACSwPLUMED/bin/mdrun_mpi

$which plumed
/vega/cheme/users/sm3334/lib/plumed/src/lib/plumed

openmpi was installed by the admin and the location is:
$which mpiexec
/usr/local/openmpi/bin/mpiexec

Another thing I noticed: each time I login I have to do "source .bashrc"

Here are the contents of my job script:

#!/bin/sh

# Directives
#PBS -N EQUIL1ns
#PBS -W group_list=yeticheme
#PBS -l nodes=2:ppn=16,walltime=01:30:00,mem=80mb
#PBS -M sm3...@columbia.edu
#PBS -m abe
#PBS -V

# Set output and error directories
#PBS -o localhost:/vega/cheme/users/sm3334/LC3/LC3-equil/1node/
#PBS -e localhost:/vega/cheme/users/sm3334/LC3/LC3-equil/1node/

cd $PBS_O_WORKDIR

module load openmpi/1.6.5-no-ib
mpiexec -n 32 mdrun_mpi -v -deffnm equil_01
# End of script


thanks in advance for your help,
Shyno

[Shyno Mathew]

Hi all,

I was able to fix the error, once I specified the LD_LIBRARY_PATH correctly, simulations are running fine.

I have few other questions related to the tutorial here: 

https://github.com/GiovanniBussi/plumed2/blob/v2.1-hrex/user-doc/tutorials/hrex.txt

After making different topology files, the .mdp file for creating different .tpr files are the same, correct?

Although below it says grompp$i.mdp, the contents of grompp0.mdp and grompp1.mdp and all other .mdp files are the same? or this is not the case?

grompp_mpi_d -maxwarn 1 -o topol$i.tpr -f grompp$i.mdp -p topol$i.top

thanks,

Shyno

On Saturday, June 20, 2015 at 7:50:32 PM UTC-4, Shyno Mathew wrote:

Dear all,

...

[Shyno Mathew]

Hi all,

1.     My previous question was about different .mdp files for creating different .tpr files corresponding to different topology files as mentioned in the tutorial:

grompp_mpi_d  -maxwarn 1 -o topol$i.tpr -f grompp$i.mdp -p topol$i.top

The .mdp files are different for only the ‘ref_t’ parameter? For eg. in my case for lamba=0, ref_t for protein should be 550 K (highest temperature I am planning to use). Is this correct?

tcoupl          = Nose-Hoover              
tc-grps         = Protein CLA_Water    
tau_t           = 0.5 0.5     
ref_t           = 550 300     

2.      The initial configuration I am using to set up REST2 is an equilibrated system of ~ 100 ns. Is there any rule of thumb here?

 

 

Thanks,

Shyno

...

[Giovanni Bussi]

Hi.

On Fri, Jun 26, 2015 at 5:09 PM, Shyno Mathew <shyno...@gmail.com> wrote:

Hi all,

1.     My previous question was about different .mdp files for creating different .tpr files corresponding to different topology files as mentioned in the tutorial:

grompp_mpi_d  -maxwarn 1 -o topol$i.tpr -f grompp$i.mdp -p topol$i.top

The .mdp files are different for only the ‘ref_t’ parameter? For eg. in my case for lamba=0, ref_t for protein should be 550 K (highest temperature I am planning to use). Is this correct?

tcoupl          = Nose-Hoover              
tc-grps         = Protein CLA_Water    
tau_t           = 0.5 0.5     
ref_t           = 550 300     

No. If you scale the force field (with partial_tempering script) you should keep temperature=300 in both replicas. 

2.      The initial configuration I am using to set up REST2 is an equilibrated system of ~ 100 ns. Is there any rule of thumb here?

No rule. Just try

Giovanni 

--

You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/ad6335d8-ba71-4507-aa28-a17798bc7ebf%40googlegroups.com.

[Shyno Mathew]

great, thanks so much Prof. Bussi!

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/0utHdmEvpnM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZKJoKyxVjfj344W%2BkUJSR9hkyZVKa%2BMZC29%2BAO0oGmXUQ%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--

Sincerely
Shyno Mathew

[Shyno Mathew]

Hi Prof. Bussi,

Just to confirm: so the .mdp files for different .tpr files (corresponding to different topology files) are exactly the same?

My simulation is crashing after few attempts of exchanging replicas with the following error:

Too many LINCS warnings

thanks,

Shyno

--

Sincerely
Shyno Mathew

[Giovanni Bussi]

LINCS warning typically is just a symptom of system explosion.

Try to run first without exchanges (-replex 1000000000).

Giovanni

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CALD%3DbpdGpEJr9y-Mpp3yP85TpFzph%2BoaGjAqDTQKXXkH5nbByQ%40mail.gmail.com.

[Shyno Mathew]

Hi Prof. Bussi,

Thanks so much for your reply,

1.      I was able to fix the issue of system crashing after few exchanges. Once I remove “-np $nrep” the system ran without any lincs warnings!

Initially I was using the following syntax as mentioned in the tutorial:

mpirun -np $nrep mdrun_mpi_d -v -plumed plumed.dat -multi $nrep -replex 100 -nsteps 15000000 –hrex

The following works:

mpirun mdrun_mpi -v -deffnm topol -plumed plumed.dat -multi 4 -replex 1000 -nsteps 500000 –hrex

But looks like no replica got exchanged, following are the last few lines of any of the 4 log files:

Replica exchange statistics
Repl  499 attempts, 250 odd, 249 even
Repl  average probabilities:
Repl     0    1    2    3
Repl      .00  .00  .00
Repl  number of exchanges:
Repl     0    1    2    3
Repl        0    0    0
Repl  average number of exchanges:
Repl     0    1    2    3
Repl      .00  .00  .00
Repl                Empirical Transition Matrix
Repl       1       2       3       4
Repl  1.0000  0.0000  0.0000  0.0000  0
Repl  0.0000  1.0000  0.0000  0.0000  1
Repl  0.0000  0.0000  1.0000  0.0000  2
Repl  0.0000  0.0000  0.0000  1.0000  3

2.    Also, I have another question:

I am using an equilibrated system (ran for 100 ns at 300 K) as my starting configuration. Before I attempt exchanging the replicas at different lambdas (effective temperatures), I need to equilibrate them at corresponding lambda values?

thanks,

Shyno

...

[Shyno Mathew]

Please see the attached .mdp file, I am using the same mdp file for all the 4 replicas, just naming the files as equil0.mdp, equil1.mp etc.

thanks,
Shyno

...

[Giovanni Bussi]

On Wed, Jul 1, 2015 at 6:24 PM, Shyno Mathew <shyno...@gmail.com> wrote:

Hi Prof. Bussi,

Thanks so much for your reply,

1.      I was able to fix the issue of system crashing after few exchanges. Once I remove “-np $nrep” the system ran without any lincs warnings!

This is very strange. Please familiarize with normal parallel tempering (as available in gromacs without any special patch) first.

 

Initially I was using the following syntax as mentioned in the tutorial:

mpirun -np $nrep mdrun_mpi_d -v -plumed plumed.dat -multi $nrep -replex 100 -nsteps 15000000 –hrex

The following works:

mpirun mdrun_mpi -v -deffnm topol -plumed plumed.dat -multi 4 -replex 1000 -nsteps 500000 –hrex

But looks like no replica got exchanged, following are the last few lines of any of the 4 log files:

Replica exchange statistics
Repl  499 attempts, 250 odd, 249 even
Repl  average probabilities:
Repl     0    1    2    3
Repl      .00  .00  .00
Repl  number of exchanges:
Repl     0    1    2    3
Repl        0    0    0
Repl  average number of exchanges:
Repl     0    1    2    3
Repl      .00  .00  .00
Repl                Empirical Transition Matrix
Repl       1       2       3       4
Repl  1.0000  0.0000  0.0000  0.0000  0
Repl  0.0000  1.0000  0.0000  0.0000  1
Repl  0.0000  0.0000  1.0000  0.0000  2
Repl  0.0000  0.0000  0.0000  1.0000  3

2.    Also, I have another question:

I am using an equilibrated system (ran for 100 ns at 300 K) as my starting configuration. Before I attempt exchanging the replicas at different lambdas (effective temperatures), I need to equilibrate them at corresponding lambda values?

This is suggested but not compulsory.

Giovanni

 

--

You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/7ddcbccd-282c-4f82-aa2a-a250367b7f1a%40googlegroups.com.

[Shyno Mathew]

ok, thanks again Prof. Bussi

Or is it because I am doing NPT?

thanks,

Shyno

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/0utHdmEvpnM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZL2KcUQwYSxtfQtsmM5PVGhErr-2AmABJt81K4ojwLSEA%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--

Sincerely
Shyno Mathew

[Shyno Mathew]

Dear all,

Once I increase the number of replicas, I was able to do both REST2 and REMD. I am trying to decide the number of replicas, I read the acceptance probability has to be greater than 30%.

I did a short simulation of 0.12 ns with an exchange interval of 2000.

1.      Here are the highest probabilities I got for different number of replicas:

16 replicas, highest average probability = 14%

32 replicas, highest average probability = 16%

48 replicas, highest average probability = 18%

I should be looking for the number of replicas that gives the best average probability for each replica, not just the highest probability?

I guess with longer runs the probability will improve?

 

16 replicas, total steps= 60,000 (0.12 ns), exchange interval= 2000
Replica exchange statistics
Repl  29 attempts, 15 odd, 14 even
Repl  average probabilities:
Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl      .00  .00  .05  .07  .00  .00  .00  .14  .00  .00  .05  .00  .00  .00  .00
Repl  number of exchanges:
Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl        0    0    1    1    0    0    0    2    0    0    1    0    1    0    0

Repl  average number of exchanges:

Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl      .00  .00  .07  .07  .00  .00  .00  .14  .00  .00  .07  .00  .07  .00  .00

32 replicas, total steps= 60,000 (0.12 ns), exchange interval= 2000
Replica exchange statistics
Repl  29 attempts, 15 odd, 14 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
Repl      .06  .01  .07  .00  .00  .04  .00  .07  .10  .01  .09  .04  .00  .07  .02  .00  .02  .08  .14  .02  .00  .04  .20  .02  .04  .02  .16  .08  .01  .00  .03
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   26   27   28   29   30   31
Repl        1    0    1    0    0    0    0    1    2    0    1    0    0    1    0    0    1    1    2    0    0    1    4    0    1    0    3    1    0    0    1

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   26   27   28   29   30   31
Repl      .07  .00  .07  .00  .00  .00  .00  .07  .13  .00  .07  .00  .00  .07  .00  .00  .07  .07  .13  .00  .00  .07  .27  .00  .07  .00  .20  .07  .00  .00  .07

48 replicas, total steps= 60,000 (0.12 ns), exchange interval= 2000
Replica exchange statistics
Repl  29 attempts, 15 odd, 14 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
Repl      .16  .00  .00  .06  .06  .15  .00  .00  .09  .03  .00  .02  .01  .14  .04  .07  .06  .00  .04  .07  .04  .07  .07  .07  .00  .01  .07  .00  .08  .07  .07  .04  .07  .01  .18  .07  .05  .13  .15  .00  .12  .06  .06  .05  .08  .07  .07
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   26   27   28   29   30   31   32   33   34   35   36   37   38   39   40   41   42   43   44   45   46   47
Repl        4    0    0    0    1    2    0    0    1    0    0    1    0    2    0    1    1    0    1    1    1    1    1    1    0    0    1    0    1    1    2    1    1    0    3    1    0    2    2    0    1    1    1    1    2    1    2

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   26   27   28   29   30   31   32   33   34   35   36   37   38   39   40   41   42   43   44   45   46   47
Repl      .27  .00  .00  .00  .07  .14  .00  .00  .07  .00  .00  .07  .00  .14  .00  .07  .07  .00  .07  .07  .07  .07  .07  .07  .00  .00  .07  .00  .07  .07  .13  .07  .07  .00  .20  .07  .00  .14  .13  .00  .07  .07  .07  .07  .13  .07  .13

 

2.      Here the simulation was ran 10 times longer (1.2 ns). There are more exchanges between different replicas.

 

16 replicas, total steps= 600,000 (1.2 ns), exchange interval= 2000
Replica exchange statistics
Repl  299 attempts, 150 odd, 149 even
Repl  average probabilities:
Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl      .02  .00  .01  .01  .06  .01  .02  .00  .02  .03  .02  .01  .03  .01  .03
Repl  number of exchanges:
Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl        4    0    1    3    9    1    3    0    3    6    3    1    5    3    5

Repl  average number of exchanges:

Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl      .03  .00  .01  .02  .06  .01  .02  .00  .02  .04  .02  .01  .03  .02  .03

 

3.    Here I made the exchanges less frequent, but looks like exchange interval= 2000 is better!

16 replicas, total steps= 600,000 (1.2 ns), exchange interval= 2000
Replica exchange statistics
Repl  149 attempts, 75 odd, 74 even
Repl  average probabilities:
Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl      .00  .00  .04  .02  .02  .01  .00  .00  .00  .04  .03  .03  .00  .01  .05
Repl  number of exchanges:
Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl        0    0    2    2    1    2    0    0    0    3    3    2    0    1    5

Repl  average number of exchanges:

Repl     0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15
Repl      .00  .00  .03  .03  .01  .03  .00  .00  .00  .04  .04  .03  .00  .01  .07

 

I am deciding the number of replicas by just looking at the average exchange probability. Is my thought process correct?

Thanks in advance for your help.

Thanks,

Shyno

...

[Shyno Mathew]

Hi all,

Forgot to mention in my previous post:

For the temperature range I used (300 K to 550 K), the lambda values change from 1 to 0.545455.

Here are the exact values:

16 replicas

1, 0.960396, 0.922362, 0.885834, 0.85075, 0.817058, 0.784699, 0.753623, 0.723776, 0.695112, 0.667583, 0.641145, 0.615753, 0.591367,  0.567947, 0.545455

32 replicas
1, 0.980636, 0.961649, 0.943029, 0.92477, 0.906862, 0.889303, 0.872083, 0.855198, 0.83864, 0.8224, 0.806478, 0.790862, 0.775549, 0.760531, 0.745805, 0.731363, 0.717202, 0.703316, 0.689698, 0.676343, 0.663247, 0.650405, 0.637811, 0.625461, 0.61335, 0.601474, 0.589828, 0.578407, 0.567207, 0.556225, 0.545455

48 replicas
1, 0.987186, 0.974535, 0.96205, 0.949722, 0.937553, 0.92554, 0.913679, 0.901971, 0.890414, 0.879005, 0.92554, 0.913679, 0.901971, 0.890414, 0.879005, 0.867742, 0.856622, 0.845647, 0.83481, 0.824112, 0.813553, 0.803129, 0.792837, 0.782679, 0.772648, 0.762749, 0.752974, 0.743327, 0.733801, 0.7244, 0.715117, 0.705954, 0.696908, 0.687977, 0.679162, 0.67046, 0.661869, 0.653387, 0.645016, 0.636751, 0.628591, 0.612585, 0.604736, 0.596986, 0.589338, 0.581785, 0.574331, 0.566972, 0.559707, 0.552535, 0.545455

Also, are there any tutorials where I can see some general analysis done on REST2 trajectories?

thanks,

Shyno

...

[Shyno Mathew]

Dear all,

I think my previous post was slightly confusing. Sorry about this!

Here is what I am trying to do:

I would like to find out the number of replicas for REST2 simulation, the temperature range is 300 K to 550 K. Initially I did three very short simulations for 0.12 ns and with exchange interval =0.004 ns. The number of replicas was 16, 32, and 48. For the probability distribution of potential energy of these runs, please see slides 1, 2, 3 of the attached file. The overlap I see for P(E) for 16 replicas (slide 1) is good enough? Or should I do some more analysis to decide the number of replicas?

I also did a slightly longer simulation 1.2ns with exchange interval 0.004 ns (slide 4) and with exchange interval 0.008 ns (slide 5).

I am also including the summary info from the log files (slides 6 to 10)

I really appreciate your input

Sincerely,

Shyno

...

[Giovanni Bussi]

Hi,

just look for any tutorial on parallel tempering on the web.

Keep in mind when you run REST2 (at least, if you use the implementation I shared with -hrex flag) that wherever you read "temperature" you should replace it with "1/lambda".

Another difference could be that whereas specific heat of water is approximately temperature independent in a relatively wide range (approximately! but there are interesting papers discussing this) so that a geometric distribution of replicas is almost always a decent starting point in explicit solvent, REST2 could be quite different in this respect. In this sense, it would be better to look for a tutorial on parallel tempering in *implicit* solvent (that's expected to be closer to REST2).

Finally, notice that overlap in energy is not required in REST2. Just look at the acceptance value. This is particularly true if you have topologies with different total charge: energy likely won't overlap, but exchanges will be accepted. That's fine.

Giovanni

--

You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/830f6a4e-c4ed-45bf-a7d6-330fcc477707%40googlegroups.com.

[Shyno Mathew]

Hi Prof. Bussi,

Thanks again for your reply. I have

As mentioned in this REST2 paper http://pubs.acs.org/doi/abs/10.1021/jp204407d 

I should be plotting the probability distribution of 

Epp+1/2 ((beta0/betam)^0.5)Epw    to get decent acceptance ratio?

Please see the attached file, I am not sure the overlap with 12 replicas is enough!

My other question is regarding the .mdp file. Following is the procedure I am following

1. Starting from an equilibrated trajectory (~100ns) I am submitting another 10 ns run for different topologies, but no exchanges during this run (using an exchange frequency higher than the number of steps)

The "continuation = no" in the .mdp file, correct. It should use the boltzmann distribution initially as  the system is exploring different temperatures.

The more I think about this I am getting confused!

2. Once the 10 ns run is finished I am using the final co-ordinates to do the actual replica exchanges

thanks,

Shyno

--
You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/0utHdmEvpnM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAPLm8ZJt0vZwVwhHaX_vxjMjf-J4ppHJ_kuUcGfKoPg3v67m7Q%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--

Sincerely
Shyno Mathew

[Shyno Mathew]

Hi all,

I am trying to do REST2, similar to the tutorial:

https://github.com/GiovanniBussi/plumed2/blob/v2.1-hrex/user-doc/tutorials/hrex.txt

I have a basic question regarding the .mdp file. Following is the procedure I am doing:

Starting from an equilibrated trajectory at 300 K (~100ns) I am submitting another 10 ns run for different topologies, but no exchanges during this run (using an exchange frequency higher than the number of steps)

The "continuation = no" in the .mdp file, correct. It should use the boltzmann distribution initially as  the system is exploring different temperatures.

Or it should be "continuation = yes" as I am starting from an equilibrated trajectory at 300 K.

Only the protein is being heated, water remains at 300 K.

The more I think about this I am getting confused!

Thanks in advance for your help,

thanks,

Shyno

...

[Shyno Mathew]

Hi Prof. Bussi,

I did a short 10 ns hrex simulation with replica exchanges at every 1000 steps. The initial coordinates for this run was taken from a 10 ns equilibration for different topologies. Please take a look at the attached file.

1. When plot the temperature of protein it's around 300 K for all replicas (page 1). Is this correct? 

2. The RMSD values are quite high (pages 2,3), but I guess I will see lower values as I run the simulations longer

thanks in advance for your help.

Shyno

--

You received this message because you are subscribed to a topic in the Google Groups "PLUMED users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plumed-users/0utHdmEvpnM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/cc6d0084-1362-48af-9723-68e2bb2da683%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Sincerely
Shyno Mathew

[shalo...@gmail.com]

If I do a quick ramachandran plot over all the residues, here is what I am seeing for the highest and lowest lambdas

thanks again for your help

Shyno

...

[Shyno Mathew]

If I do a quick ramachandran plot over all the residues, here is what I am seeing for the highest and lowest lambdas

thanks again for your help

Shyno

[Giovanni Bussi]

On Thu, Aug 6, 2015 at 11:04 PM, Shyno Mathew <shyno...@gmail.com> wrote:

Hi Prof. Bussi,

Thanks again for your reply. I have

As mentioned in this REST2 paper http://pubs.acs.org/doi/abs/10.1021/jp204407d 

I should be plotting the probability distribution of 

Epp+1/2 ((beta0/betam)^0.5)Epw    to get decent acceptance ratio?

Please see the attached file, I am not sure the overlap with 12 replicas is enough!

Just check the acceptance directly. This is simpler.

 

My other question is regarding the .mdp file. Following is the procedure I am following

1. Starting from an equilibrated trajectory (~100ns) I am submitting another 10 ns run for different topologies, but no exchanges during this run (using an exchange frequency higher than the number of steps)

The "continuation = no" in the .mdp file, correct. It should use the boltzmann distribution initially as  the system is exploring different temperatures.

The more I think about this I am getting confused!

This is strictly not necessary but for sure not wrong. I also typically do something similar.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CALD%3DbpeDJJGBrxYgWf3dbJBA5YgjZ68e2KJ0m%3DZrj66ORqOk%3DQ%40mail.gmail.com.

[Giovanni Bussi]

On Mon, Aug 24, 2015 at 12:04 AM, Shyno Mathew <shyno...@gmail.com> wrote:

Hi Prof. Bussi,

I did a short 10 ns hrex simulation with replica exchanges at every 1000 steps. The initial coordinates for this run was taken from a 10 ns equilibration for different topologies. Please take a look at the attached file.

1. When plot the temperature of protein it's around 300 K for all replicas (page 1). Is this correct? 

It is. In this implementation of REST2, all replicas are simulated at the same temperature and only the force field is scaled. What matters is the lambda factor. The "effective temperature" (300/lambda) is just a number to have a physical insights on the ensemble that you are exploring. The thermostat should be set to 300 anyway.

 

2. The RMSD values are quite high (pages 2,3), but I guess I will see lower values as I run the simulations longer

This depends on the system. It is difficult to judge this without knowing exactly what you are trying to do. However, in this mailing list you are expected to get technical support on using the implemented technique. If you need help in interpreting your results perhaps you have to collaborate with someone familiar with the methodologies.

Giovanni

 

--

You received this message because you are subscribed to the Google Groups "PLUMED users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.

To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CALD%3DbpfoJ9vVAE-LtYQGO9DGMfupXRRRsYQLr73Y0nOmYxKcoQ%40mail.gmail.com.

[Shyno Mathew]

Hi Prof. Bussi,

Thanks so much for your help.

Thanks again for the suggestion, I am going to ask my advisor about the collaboration.

Sincerely,
Shyno

...

[Shyno Mathew]

Hi Prof. Bussi,

If possible, could you answer the following question.

In your paper "Hamiltoninan replica exchange in GROMACS: a flexible implementation" it says on page 381, "exchanges are attempted every 100 steps". Here 100 steps means 0.2 ps ? I am trying to figure out the a good exchange interval. I also read in gromacs documentation the interval should not be under 1 ps for protein systems!

http://www.gromacs.org/Documentation/How-tos/REMD 

I tried the intervals 2 ps, 0.2 ps.

thanks,

Shyno

...

[Giovanni Bussi]

Hi,

yes in my paper it was 200 fs.

In my opinion there is no reason not to use an interval < 1ps, except from the fact that the simulations slow down a bit. Aside from this technical issue, the shorter stride is the better. This is one paper explaining this (but there might be others):

http://scitation.aip.org/content/aip/journal/jcp/128/2/10.1063/1.2816560

Giovanni

--

You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/504ddf9d-9597-4b9d-9378-5f7e2350f50f%40googlegroups.com.