[gmx-users] QUERY IN GRO FILE

[Neha Gupta]

Hi gromacs users,


I have a doubt in this command,

grompp -f em.mdp -c conf.gro -p topol.top -o em


mdrun -v -deffnm em


These commands generate em.gro, em.tpr, em.trr, em.log file.


There is another command which is


mdrun -s em.tpr

which generates confout.gro

Does em.gro and confout.gro are one and the same?

What are the technical differences?


Thanks,

Neha
-- 
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-requ...@gromacs.org.

[gmx-users] terminated abnormally, message of GridMAT-MD tool

[Мижээ Батсайхан]

Dear Justin,

I am using GridMAT-MD tool for analyses of heterogeneous membranes.
I generated trajectory.gro only a group which contains a peptide and
lipids.

I got following
Thread 1 terminated abnormally: Illegal division by zero at
GridMAT-MD-parallel.pl line 501.

What is the meaning of the message of the tool?


Best regards,
Mijee
-- 
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-requ...@gromacs.org.

Re: [gmx-users] Box dimensions in the gro file are not read by GROMACS

[Sajeewa Pemasinghe]

Hi Justin,

I am really sorry for taking your time. I found my mistake. Earlier this
error was not there because I was applying my umbrella potential along x
axis only which has almost 120 A. Now I am applying umbrella potential
along both x and y axes. That is causing the problem as the y axis
dimension is not big enough. I will have to equilibrate the system in a
bigger box with a larger y dimension and run the umbrella sampling.

Thank you very much for your time.

Best,

Sajeewa

On Wed, Mar 29, 2017 at 11:26 PM, Sajeewa Pemasinghe 
wrote:

> So if the actual distance is 2.3 nm why does GROMACS measure it as 3.58
> nm ?
>
> On Wed, Mar 29, 2017 at 11:11 PM, Sajeewa Pemasinghe 
> wrote:
>
>> The coordinates of group1 center of mass are 44.466, 23.909,35.035 and
>> that of group 2 are 67.467,23.994,34.088 as calculated in VMD . So the
>> distance is 23 A .
>>
>> On Wed, Mar 29, 2017 at 11:05 PM, Sajeewa Pemasinghe > > wrote:
>>
>>> Hi Justin,
>>>
>>> Thank you for the prompt reply. I asked a second question correcting the
>>> value of 23.0 nm. The distance value is 2.3 nm.
>>>
>>> On Wed, Mar 29, 2017 at 10:57 PM, Justin Lemkul  wrote:
>>>


 On 3/29/17 2:55 PM, Sajeewa Pemasinghe wrote:

> Hi everyone,
>
> I have run an equilibration of two linear molecules in a water box with
> dimensions 117A x 54A x 70A for 150 ns. I am using the equilibrated
> structures (not the final structure) in the final nanoseconds as
> starting
> structures for an umbrella sampling calculation. The problem is that
> as I
> only have a checkpoint (cpt) file for the final structure I have to
> start
> these umbrella sampling calculations without a cpt file. So my inputs
> to
> grompp are just the .gro files. During the umbrella sampling I get the
> following error
>
> Distance between pull groups 1 and 2 (3.58 nm) is larger than 0.49
> times the box size (2.673675).
> You might want to consider using "pull-geometry = direction-periodic"
> instead.
>
> Although gromacs says that the distance between the pull groups 1 and
> 2 are
> 3.58 nm, the actual distance in the snapshot is 23.0 nm. This was
> confirmed by calculating in vmd as well.
> The above error also says that the box size is 2.673675 nm. But at the
> bottom of the input gro file it shows the dimensions as 11.68645
> 5.34804   6.95245. So I cannot understand why it doesn't read those
> dimensions.
>
>
 It is.  Half of 5.34804 is the value being reported in the error.  You
 also can't possibly have a distance of 23.0 nm if the longest box vector is
 half that value.

 -Justin

 --
 ==

 Justin A. Lemkul, Ph.D.
 Ruth L. Kirschstein NRSA Postdoctoral Fellow

 Department of Pharmaceutical Sciences
 School of Pharmacy
 Health Sciences Facility II, Room 629
 University of Maryland, Baltimore
 20 Penn St.
 Baltimore, MD 21201

 jalem...@outerbanks.umaryland.edu | (410) 706-7441
 http://mackerell.umaryland.edu/~jalemkul

 ==
 --
 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-requ...@gromacs.org.

>>>
>>>
>>
>
-- 
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-requ...@gromacs.org.

Re: [gmx-users] Box dimensions in the gro file are not read by GROMACS

[Sajeewa Pemasinghe]

So if the actual distance is 2.3 nm why does GROMACS measure it as 3.58
nm ?

On Wed, Mar 29, 2017 at 11:11 PM, Sajeewa Pemasinghe 
wrote:

> The coordinates of group1 center of mass are 44.466, 23.909,35.035 and
> that of group 2 are 67.467,23.994,34.088 as calculated in VMD . So the
> distance is 23 A .
>
> On Wed, Mar 29, 2017 at 11:05 PM, Sajeewa Pemasinghe 
> wrote:
>
>> Hi Justin,
>>
>> Thank you for the prompt reply. I asked a second question correcting the
>> value of 23.0 nm. The distance value is 2.3 nm.
>>
>> On Wed, Mar 29, 2017 at 10:57 PM, Justin Lemkul  wrote:
>>
>>>
>>>
>>> On 3/29/17 2:55 PM, Sajeewa Pemasinghe wrote:
>>>
 Hi everyone,

 I have run an equilibration of two linear molecules in a water box with
 dimensions 117A x 54A x 70A for 150 ns. I am using the equilibrated
 structures (not the final structure) in the final nanoseconds as
 starting
 structures for an umbrella sampling calculation. The problem is that as
 I
 only have a checkpoint (cpt) file for the final structure I have to
 start
 these umbrella sampling calculations without a cpt file. So my inputs to
 grompp are just the .gro files. During the umbrella sampling I get the
 following error

 Distance between pull groups 1 and 2 (3.58 nm) is larger than 0.49
 times the box size (2.673675).
 You might want to consider using "pull-geometry = direction-periodic"
 instead.

 Although gromacs says that the distance between the pull groups 1 and 2
 are
 3.58 nm, the actual distance in the snapshot is 23.0 nm. This was
 confirmed by calculating in vmd as well.
 The above error also says that the box size is 2.673675 nm. But at the
 bottom of the input gro file it shows the dimensions as 11.68645
 5.34804   6.95245. So I cannot understand why it doesn't read those
 dimensions.


>>> It is.  Half of 5.34804 is the value being reported in the error.  You
>>> also can't possibly have a distance of 23.0 nm if the longest box vector is
>>> half that value.
>>>
>>> -Justin
>>>
>>> --
>>> ==
>>>
>>> Justin A. Lemkul, Ph.D.
>>> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>>>
>>> Department of Pharmaceutical Sciences
>>> School of Pharmacy
>>> Health Sciences Facility II, Room 629
>>> University of Maryland, Baltimore
>>> 20 Penn St.
>>> Baltimore, MD 21201
>>>
>>> jalem...@outerbanks.umaryland.edu | (410) 706-7441
>>> http://mackerell.umaryland.edu/~jalemkul
>>>
>>> ==
>>> --
>>> 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-requ...@gromacs.org.
>>>
>>
>>
>
-- 
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-requ...@gromacs.org.

Re: [gmx-users] Box dimensions in the gro file are not read by GROMACS

[Sajeewa Pemasinghe]

The coordinates of group1 center of mass are 44.466, 23.909,35.035 and that
of group 2 are 67.467,23.994,34.088 as calculated in VMD . So the distance
is 23 A .

On Wed, Mar 29, 2017 at 11:05 PM, Sajeewa Pemasinghe 
wrote:

> Hi Justin,
>
> Thank you for the prompt reply. I asked a second question correcting the
> value of 23.0 nm. The distance value is 2.3 nm.
>
> On Wed, Mar 29, 2017 at 10:57 PM, Justin Lemkul  wrote:
>
>>
>>
>> On 3/29/17 2:55 PM, Sajeewa Pemasinghe wrote:
>>
>>> Hi everyone,
>>>
>>> I have run an equilibration of two linear molecules in a water box with
>>> dimensions 117A x 54A x 70A for 150 ns. I am using the equilibrated
>>> structures (not the final structure) in the final nanoseconds as starting
>>> structures for an umbrella sampling calculation. The problem is that as I
>>> only have a checkpoint (cpt) file for the final structure I have to start
>>> these umbrella sampling calculations without a cpt file. So my inputs to
>>> grompp are just the .gro files. During the umbrella sampling I get the
>>> following error
>>>
>>> Distance between pull groups 1 and 2 (3.58 nm) is larger than 0.49
>>> times the box size (2.673675).
>>> You might want to consider using "pull-geometry = direction-periodic"
>>> instead.
>>>
>>> Although gromacs says that the distance between the pull groups 1 and 2
>>> are
>>> 3.58 nm, the actual distance in the snapshot is 23.0 nm. This was
>>> confirmed by calculating in vmd as well.
>>> The above error also says that the box size is 2.673675 nm. But at the
>>> bottom of the input gro file it shows the dimensions as 11.68645
>>> 5.34804   6.95245. So I cannot understand why it doesn't read those
>>> dimensions.
>>>
>>>
>> It is.  Half of 5.34804 is the value being reported in the error.  You
>> also can't possibly have a distance of 23.0 nm if the longest box vector is
>> half that value.
>>
>> -Justin
>>
>> --
>> ==
>>
>> Justin A. Lemkul, Ph.D.
>> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>>
>> Department of Pharmaceutical Sciences
>> School of Pharmacy
>> Health Sciences Facility II, Room 629
>> University of Maryland, Baltimore
>> 20 Penn St.
>> Baltimore, MD 21201
>>
>> jalem...@outerbanks.umaryland.edu | (410) 706-7441
>> http://mackerell.umaryland.edu/~jalemkul
>>
>> ==
>> --
>> 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-requ...@gromacs.org.
>>
>
>
-- 
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-requ...@gromacs.org.

Re: [gmx-users] Box dimensions in the gro file are not read by GROMACS

[Sajeewa Pemasinghe]

Hi Justin,

Thank you for the prompt reply. I asked a second question correcting the
value of 23.0 nm. The distance value is 2.3 nm.

On Wed, Mar 29, 2017 at 10:57 PM, Justin Lemkul  wrote:

>
>
> On 3/29/17 2:55 PM, Sajeewa Pemasinghe wrote:
>
>> Hi everyone,
>>
>> I have run an equilibration of two linear molecules in a water box with
>> dimensions 117A x 54A x 70A for 150 ns. I am using the equilibrated
>> structures (not the final structure) in the final nanoseconds as starting
>> structures for an umbrella sampling calculation. The problem is that as I
>> only have a checkpoint (cpt) file for the final structure I have to start
>> these umbrella sampling calculations without a cpt file. So my inputs to
>> grompp are just the .gro files. During the umbrella sampling I get the
>> following error
>>
>> Distance between pull groups 1 and 2 (3.58 nm) is larger than 0.49
>> times the box size (2.673675).
>> You might want to consider using "pull-geometry = direction-periodic"
>> instead.
>>
>> Although gromacs says that the distance between the pull groups 1 and 2
>> are
>> 3.58 nm, the actual distance in the snapshot is 23.0 nm. This was
>> confirmed by calculating in vmd as well.
>> The above error also says that the box size is 2.673675 nm. But at the
>> bottom of the input gro file it shows the dimensions as 11.68645
>> 5.34804   6.95245. So I cannot understand why it doesn't read those
>> dimensions.
>>
>>
> It is.  Half of 5.34804 is the value being reported in the error.  You
> also can't possibly have a distance of 23.0 nm if the longest box vector is
> half that value.
>
> -Justin
>
> --
> ==
>
> Justin A. Lemkul, Ph.D.
> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>
> Department of Pharmaceutical Sciences
> School of Pharmacy
> Health Sciences Facility II, Room 629
> University of Maryland, Baltimore
> 20 Penn St.
> Baltimore, MD 21201
>
> jalem...@outerbanks.umaryland.edu | (410) 706-7441
> http://mackerell.umaryland.edu/~jalemkul
>
> ==
> --
> 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-requ...@gromacs.org.
>
-- 
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-requ...@gromacs.org.

Re: [gmx-users] Box dimensions in the gro file are not read by GROMACS

[Justin Lemkul]

On 3/29/17 2:55 PM, Sajeewa Pemasinghe wrote:

Hi everyone,

I have run an equilibration of two linear molecules in a water box with
dimensions 117A x 54A x 70A for 150 ns. I am using the equilibrated
structures (not the final structure) in the final nanoseconds as starting
structures for an umbrella sampling calculation. The problem is that as I
only have a checkpoint (cpt) file for the final structure I have to start
these umbrella sampling calculations without a cpt file. So my inputs to
grompp are just the .gro files. During the umbrella sampling I get the
following error

Distance between pull groups 1 and 2 (3.58 nm) is larger than 0.49
times the box size (2.673675).
You might want to consider using "pull-geometry = direction-periodic"
instead.

Although gromacs says that the distance between the pull groups 1 and 2 are
3.58 nm, the actual distance in the snapshot is 23.0 nm. This was
confirmed by calculating in vmd as well.
The above error also says that the box size is 2.673675 nm. But at the
bottom of the input gro file it shows the dimensions as 11.68645
5.34804   6.95245. So I cannot understand why it doesn't read those
dimensions.



It is.  Half of 5.34804 is the value being reported in the error.  You also 
can't possibly have a distance of 23.0 nm if the longest box vector is half that 
value.


-Justin

--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalem...@outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==
--
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-requ...@gromacs.org.

[gmx-users] Box dimensions in the gro file are not read by GROMACS (correcting an error in the previous question)

[Sajeewa Pemasinghe]

Hi everyone,

I have run an equilibration of two linear molecules in a water box with
dimensions 117A x 54A x 70A for 150 ns. I am using the equilibrated
structures (not the final structure) in the final nanoseconds as starting
structures for an umbrella sampling calculation. The problem is that as I
only have a checkpoint (cpt) file for the final structure I have to start
these umbrella sampling calculations without a cpt file. So my inputs to
grompp are just the .gro files. During the umbrella sampling I get the
following error

Distance between pull groups 1 and 2 (3.58 nm) is larger than 0.49
times the box size (2.673675).
You might want to consider using "pull-geometry = direction-periodic"
instead.

Although gromacs says that the distance between the pull groups 1 and 2 are
3.58 nm, the actual distance in the snapshot is *2.3 *nm. This was
confirmed by calculating in vmd as well.
The above error also says that the box size is 2.673675 nm. But at the
bottom of the input gro file it shows the dimensions as 11.68645
5.34804   6.95245. So I cannot understand why it doesn't read those
dimensions.

I am doing this umbrella sampling in the NPT ensemble. As the gromacs
manual suggests "no-pressure scaling" for "pull-geometry =
direction-periodic" I cannot use it either.

I would highly appreciate some help with this problem.

Thank you very much
-- 
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-requ...@gromacs.org.

[gmx-users] Box dimensions in the gro file are not read by GROMACS

[Sajeewa Pemasinghe]

Hi everyone,

I have run an equilibration of two linear molecules in a water box with
dimensions 117A x 54A x 70A for 150 ns. I am using the equilibrated
structures (not the final structure) in the final nanoseconds as starting
structures for an umbrella sampling calculation. The problem is that as I
only have a checkpoint (cpt) file for the final structure I have to start
these umbrella sampling calculations without a cpt file. So my inputs to
grompp are just the .gro files. During the umbrella sampling I get the
following error

Distance between pull groups 1 and 2 (3.58 nm) is larger than 0.49
times the box size (2.673675).
You might want to consider using "pull-geometry = direction-periodic"
instead.

Although gromacs says that the distance between the pull groups 1 and 2 are
3.58 nm, the actual distance in the snapshot is 23.0 nm. This was
confirmed by calculating in vmd as well.
The above error also says that the box size is 2.673675 nm. But at the
bottom of the input gro file it shows the dimensions as 11.68645
5.34804   6.95245. So I cannot understand why it doesn't read those
dimensions.

I am doing this umbrella sampling in the NPT ensemble. As the gromacs
manual suggests "no-pressure scaling" for "pull-geometry =
direction-periodic" I cannot use it either.

I would highly appreciate some help with this problem.

Thank you very much

Sajeewa
-- 
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-requ...@gromacs.org.

Re: [gmx-users] Format for predefined positions insert-molecules

[Justin Lemkul]

On 3/29/17 12:52 PM, Svetomir Hitov wrote:

Hello,

I have been going through the documentation and mailing list for the last
couple of days, but I cannot find anywhere, what the format for the
predefined insertion trial positions should be. Insert-molecules should use


I'd start with the last paragraph of the help information:

"Alternatively, the molecules can be inserted only at positions defined in
positions.dat (-ip). That file should have 3 columns (x,y,z), that give the
displacements compared to the input molecule position (-ci). Hence, if that
file should contain the absolute positions, the molecule must be centered on
(0,0,0) before using gmx insert-molecules (e.g. from gmx editconf -center).
Comments in that file starting with # are ignored."

-Justin


that file to place molecules inside the box in certain places. I have tried
plain text file with several coordinates. Insert-molecules is not giving any
error but yet it is inserting molecules randomly. I want to make it insert
molecules just in one half of the box.
Thank you in advance!



--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalem...@outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==
--
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-requ...@gromacs.org.

Re: [gmx-users] mdrun error

[Justin Lemkul]

On 3/29/17 11:49 AM, amitbe...@chemeng.iisc.ernet.in wrote:

Hello,
After equilibrating my system for 100 ns( during which there was restraint
on the protein letting the lipids to equilibrate), when I am trying to run
my unrestrained run I am getting the following errors:
1. "bonds that rotated more than 30 degree"
2. "1 particles communicated to PME rank 0 are more than 2/3 times the
cut-off out of the domain decomposition cell of their charge group in
dimension x"
3. "Listed nonbonded interaction between particles 71916 and 71920
at distance 5.535 which is larger than the table limit 2.215 nm."

What might be the reason for this error. I thought 100ns is enough for
equilibration.



Something is likely topologically unsound.  The restraints were probably masking 
an underlying issue (100 ns is way longer than one normally runs for restrained 
dynamics, FWIW).


-Justin

--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalem...@outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==
--
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-requ...@gromacs.org.

Re: [gmx-users] Replica Exchange for surface tension

[Justin Lemkul]

On 3/29/17 11:29 AM, gozde ergin wrote:

Justin what would you suggest if I simulate the system under constant surface 
tension since I know the correct surface tension from experiments.
Can I get some information about the force field quality if I compare the 
conformation of the surface under constant surface tension (N\gammaT) and under 
NVT ensemble?



Possibly, but I would still look at the persistence of any interactions in the 
mixed system before I invested more time in another (potentially long) 
simulation.  There should be some clues already available to you.


-Justin


On 29 Mar 2017, at 15:21, gozde ergin  wrote:

Exactly SDS is CHARMM36 but cocobetaine not.


On 29 Mar 2017, at 15:18, Justin Lemkul > wrote:



On 3/29/17 9:16 AM, gozde ergin wrote:

Are you sure it is not CHARMM36 because it is written:

;;
;; Generated by CHARMM-GUI (http://www.charmm-gui.org 
) v1.7
;;
;; psf2itp.py
;;
;; Correspondance:
;; j712l...@ku.edu  or won...@ku.edu 

;;
;; CHARMM36 FF in GROMACS format
;;

on the top of the .itp files.



OK, I was thinking too quickly, the workflow goes like this: If a molecule is identified 
as part of the existing force field, it is used.  If it is not, then CGenFF is called.  
So your SDS may be CHARMM36 but your cocobetaine is not. You can check against the CHARMM 
force field files and also the atom types themselves.  If the second character is 
"G" then it is CGenFF, otherwise it is stock CHARMM.

Do not rely on header comments; those are often boilerplate.  Check the 
parameters themselves.

-Justin





On 29 Mar 2017, at 15:06, Justin Lemkul > wrote:



On 3/29/17 4:15 AM, gozde ergin wrote:

I used CHARMM GUI Input Generator tool both for SDS and cocobetaine.


For the sake of clarity, you should understand that you are *not* then using 
CHARMM36.  CHARMM-GUI calls the CGenFF program to parametrize the molecule, so 
you are using CGenFF parameters, which are compatible with CHARMM by design, 
but are not the standard CHARMM36 parameters.  Thus your SDS model will differ 
from the standard one that we have in the lipids/detergents file.  You should 
compare it to make sure it is reasonable.  A general force field will sacrifice 
some accuracy in the name of generality and transferability.


So basically the reason of not getting correct surface tension is bad 
parametrisation of force field.
Then I should not spend time on simulation the replica exchange.


Indeed.  Look at the things I mentioned before.  If the model doesn't work 
well, then you have to get into the details of why.  Burning through more time 
hoping it will work out and expensive stuff like REMD is just a waste.

-Justin

--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalem...@outerbanks.umaryland.edu  | 
(410) 706-7441
http://mackerell.umaryland.edu/~jalemkul 


==
--
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-requ...@gromacs.org.




--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalem...@outerbanks.umaryland.edu  | 
(410) 706-7441
http://mackerell.umaryland.edu/~jalemkul 


==
--
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-requ...@gromacs.org .






--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical 

[gmx-users] Format for predefined positions insert-molecules

[Svetomir Hitov]

Hello,

I have been going through the documentation and mailing list for the last
couple of days, but I cannot find anywhere, what the format for the
predefined insertion trial positions should be. Insert-molecules should use
that file to place molecules inside the box in certain places. I have tried
plain text file with several coordinates. Insert-molecules is not giving any
error but yet it is inserting molecules randomly. I want to make it insert
molecules just in one half of the box.
Thank you in advance! 

-- 
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-requ...@gromacs.org.

[gmx-users] mdrun error

[amitbehra]

Hello,
After equilibrating my system for 100 ns( during which there was restraint
on the protein letting the lipids to equilibrate), when I am trying to run
my unrestrained run I am getting the following errors:
1. "bonds that rotated more than 30 degree"
2. "1 particles communicated to PME rank 0 are more than 2/3 times the
cut-off out of the domain decomposition cell of their charge group in
dimension x"
3. "Listed nonbonded interaction between particles 71916 and 71920
at distance 5.535 which is larger than the table limit 2.215 nm."

What might be the reason for this error. I thought 100ns is enough for
equilibration.

Thanks,
Amit Behera
Research Scholar,
IISc.

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

-- 
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-requ...@gromacs.org.

Re: [gmx-users] Replica Exchange for surface tension

[gozde ergin]

Justin what would you suggest if I simulate the system under constant surface 
tension since I know the correct surface tension from experiments.
Can I get some information about the force field quality if I compare the 
conformation of the surface under constant surface tension (N\gammaT) and under 
NVT ensemble?
 
> On 29 Mar 2017, at 15:21, gozde ergin  wrote:
> 
> Exactly SDS is CHARMM36 but cocobetaine not.
> 
>> On 29 Mar 2017, at 15:18, Justin Lemkul > > wrote:
>> 
>> 
>> 
>> On 3/29/17 9:16 AM, gozde ergin wrote:
>>> Are you sure it is not CHARMM36 because it is written:
>>> 
>>> ;;
>>> ;; Generated by CHARMM-GUI (http://www.charmm-gui.org 
>>> ) v1.7
>>> ;;
>>> ;; psf2itp.py
>>> ;;
>>> ;; Correspondance:
>>> ;; j712l...@ku.edu  or won...@ku.edu 
>>> 
>>> ;;
>>> ;; CHARMM36 FF in GROMACS format
>>> ;;
>>> 
>>> on the top of the .itp files.
>>> 
>> 
>> OK, I was thinking too quickly, the workflow goes like this: If a molecule 
>> is identified as part of the existing force field, it is used.  If it is 
>> not, then CGenFF is called.  So your SDS may be CHARMM36 but your 
>> cocobetaine is not. You can check against the CHARMM force field files and 
>> also the atom types themselves.  If the second character is "G" then it is 
>> CGenFF, otherwise it is stock CHARMM.
>> 
>> Do not rely on header comments; those are often boilerplate.  Check the 
>> parameters themselves.
>> 
>> -Justin
>> 
>>> 
>>> 
 On 29 Mar 2017, at 15:06, Justin Lemkul > wrote:
 
 
 
 On 3/29/17 4:15 AM, gozde ergin wrote:
> I used CHARMM GUI Input Generator tool both for SDS and cocobetaine.
 
 For the sake of clarity, you should understand that you are *not* then 
 using CHARMM36.  CHARMM-GUI calls the CGenFF program to parametrize the 
 molecule, so you are using CGenFF parameters, which are compatible with 
 CHARMM by design, but are not the standard CHARMM36 parameters.  Thus your 
 SDS model will differ from the standard one that we have in the 
 lipids/detergents file.  You should compare it to make sure it is 
 reasonable.  A general force field will sacrifice some accuracy in the 
 name of generality and transferability.
 
> So basically the reason of not getting correct surface tension is bad 
> parametrisation of force field.
> Then I should not spend time on simulation the replica exchange.
 
 Indeed.  Look at the things I mentioned before.  If the model doesn't work 
 well, then you have to get into the details of why.  Burning through more 
 time hoping it will work out and expensive stuff like REMD is just a waste.
 
 -Justin
 
 --
 ==
 
 Justin A. Lemkul, Ph.D.
 Ruth L. Kirschstein NRSA Postdoctoral Fellow
 
 Department of Pharmaceutical Sciences
 School of Pharmacy
 Health Sciences Facility II, Room 629
 University of Maryland, Baltimore
 20 Penn St.
 Baltimore, MD 21201
 
 jalem...@outerbanks.umaryland.edu 
  | (410) 706-7441
 http://mackerell.umaryland.edu/~jalemkul 
 
 
 ==
 --
 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-requ...@gromacs.org.
>>> 
>> 
>> -- 
>> ==
>> 
>> Justin A. Lemkul, Ph.D.
>> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>> 
>> Department of Pharmaceutical Sciences
>> School of Pharmacy
>> Health Sciences Facility II, Room 629
>> University of Maryland, Baltimore
>> 20 Penn St.
>> Baltimore, MD 21201
>> 
>> jalem...@outerbanks.umaryland.edu  
>> | (410) 706-7441
>> http://mackerell.umaryland.edu/~jalemkul 
>> 
>> 
>> ==
>> -- 
>> 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 

Re: [gmx-users] importance of nvlink in GPU nodes?

[Christopher Neale]

thank you very much Mark and Szilard!

From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se 
 on behalf of Szilárd Páll 

Sent: 29 March 2017 10:09:05
To: Discussion list for GROMACS users
Subject: Re: [gmx-users] importance of nvlink in GPU nodes?

Note that while, as Mark said, GPU-to-GPU NVLink is not used optimized for,
when NVLink is the main CPU-to-GPU interconnect, GROMACS will certainly
transfer data across it! Currently only IBM Minsky systems have direct
CPU-to-GPU NVLink; on these transfers are ~3x faster than on PCI-E-based
systems. While transfers generally don't take much time, given that the
P100 GPUs are in general ~2.5x faster than previous generation, with PCI-E
the relative time spent in transfers will increase quite significantly.

So to answer the original question: the CPU-to-GPU NVLink can have a slight
impact on performance, but I'd expect that to be at most 10%, typically
less.

Cheers,
--
Szilárd

On Wed, Mar 29, 2017 at 2:46 PM, Mark Abraham 
wrote:

> Hi,
>
> There's no use made of NVLINK in GROMACS. The 2018 release might make some
> use of it.
>
> Mark
>
> On Wed, Mar 29, 2017 at 4:19 AM Christopher Neale <
> chris.ne...@alum.utoronto.ca> wrote:
>
> > Dear users:
> >
> > does anybody know if the presence of either CPU-to-GPU or GPU-to-GPU
> > nvlink affects the performance or efficiency of gromacs?
> >
> > Thank you,
> > Chris.
> > --
> > 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-requ...@gromacs.org.
> >
> --
> 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-requ...@gromacs.org.
>
--
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-requ...@gromacs.org.
-- 
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-requ...@gromacs.org.

Re: [gmx-users] potential energy of a system

[Saumyak Mukherjee]

Dear Mark and Justin,

Thanks a lot for the information

On 29 March 2017 at 18:33, Justin Lemkul  wrote:

>
>
> On 3/29/17 8:43 AM, Mark Abraham wrote:
>
>> Hi,
>>
>> The total potential energy has everything by definition, so I don't know
>> what you're asking.
>>
>> Additivity of the components of the potential energy is an artefact of the
>> construction of the model physics. While it is true that different
>> physical
>> phenomena underlie some of the components, the parameterization of the
>> force field did not target meaningful decomposition of the potential
>> energy. But you might be able to benefit from fortuitous cancellation
>> effects when comparing such things.
>>
>>
> Nitpick here: in CHARMM (and our polarizable model), we seek to get a
> variety of QM interaction energies explicitly correct (interactions with
> water, ions, other small molecules) so indeed there is some reliability in
> these metrics with CHARMM.  Of course, like all force fields, there is some
> inherent error cancellation but we do make an effort to get all of this
> kind of stuff right. Most other force fields, of course, do not, seeking
> instead to get overall properties correct, leaving the details to be
> debatable.  We try to get it all :)
>
> -Justin
>
> --
> ==
>
> Justin A. Lemkul, Ph.D.
> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>
> Department of Pharmaceutical Sciences
> School of Pharmacy
> Health Sciences Facility II, Room 629
> University of Maryland, Baltimore
> 20 Penn St.
> Baltimore, MD 21201
>
> jalem...@outerbanks.umaryland.edu | (410) 706-7441
> http://mackerell.umaryland.edu/~jalemkul
>
> ==
>
> --
> 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-requ...@gromacs.org.
>



-- 

*Saumyak Mukherjee*

Junior Research Fellow
Prof. Biman Bagchi's Group
Solid State and Structural Chemistry Unit
Indian Institute of Science
Bangalore - 560012

Mob : 8017292426
Alternative e-mail : saumyakmukher...@gmail.com
smukher...@sscu.iisc.ernet.in

-- 
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-requ...@gromacs.org.

Re: [gmx-users] importance of nvlink in GPU nodes?

[Szilárd Páll]

Note that while, as Mark said, GPU-to-GPU NVLink is not used optimized for,
when NVLink is the main CPU-to-GPU interconnect, GROMACS will certainly
transfer data across it! Currently only IBM Minsky systems have direct
CPU-to-GPU NVLink; on these transfers are ~3x faster than on PCI-E-based
systems. While transfers generally don't take much time, given that the
P100 GPUs are in general ~2.5x faster than previous generation, with PCI-E
the relative time spent in transfers will increase quite significantly.

So to answer the original question: the CPU-to-GPU NVLink can have a slight
impact on performance, but I'd expect that to be at most 10%, typically
less.

Cheers,
--
Szilárd

On Wed, Mar 29, 2017 at 2:46 PM, Mark Abraham 
wrote:

> Hi,
>
> There's no use made of NVLINK in GROMACS. The 2018 release might make some
> use of it.
>
> Mark
>
> On Wed, Mar 29, 2017 at 4:19 AM Christopher Neale <
> chris.ne...@alum.utoronto.ca> wrote:
>
> > Dear users:
> >
> > does anybody know if the presence of either CPU-to-GPU or GPU-to-GPU
> > nvlink affects the performance or efficiency of gromacs?
> >
> > Thank you,
> > Chris.
> > --
> > 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-requ...@gromacs.org.
> >
> --
> 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-requ...@gromacs.org.
>
-- 
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-requ...@gromacs.org.

Re: [gmx-users] Replica Exchange for surface tension

[gozde ergin]

Exactly SDS is CHARMM36 but cocobetaine not.

> On 29 Mar 2017, at 15:18, Justin Lemkul  wrote:
> 
> 
> 
> On 3/29/17 9:16 AM, gozde ergin wrote:
>> Are you sure it is not CHARMM36 because it is written:
>> 
>> ;;
>> ;; Generated by CHARMM-GUI (http://www.charmm-gui.org) v1.7
>> ;;
>> ;; psf2itp.py
>> ;;
>> ;; Correspondance:
>> ;; j712l...@ku.edu or won...@ku.edu
>> ;;
>> ;; CHARMM36 FF in GROMACS format
>> ;;
>> 
>> on the top of the .itp files.
>> 
> 
> OK, I was thinking too quickly, the workflow goes like this: If a molecule is 
> identified as part of the existing force field, it is used.  If it is not, 
> then CGenFF is called.  So your SDS may be CHARMM36 but your cocobetaine is 
> not. You can check against the CHARMM force field files and also the atom 
> types themselves.  If the second character is "G" then it is CGenFF, 
> otherwise it is stock CHARMM.
> 
> Do not rely on header comments; those are often boilerplate.  Check the 
> parameters themselves.
> 
> -Justin
> 
>> 
>> 
>>> On 29 Mar 2017, at 15:06, Justin Lemkul  wrote:
>>> 
>>> 
>>> 
>>> On 3/29/17 4:15 AM, gozde ergin wrote:
 I used CHARMM GUI Input Generator tool both for SDS and cocobetaine.
>>> 
>>> For the sake of clarity, you should understand that you are *not* then 
>>> using CHARMM36.  CHARMM-GUI calls the CGenFF program to parametrize the 
>>> molecule, so you are using CGenFF parameters, which are compatible with 
>>> CHARMM by design, but are not the standard CHARMM36 parameters.  Thus your 
>>> SDS model will differ from the standard one that we have in the 
>>> lipids/detergents file.  You should compare it to make sure it is 
>>> reasonable.  A general force field will sacrifice some accuracy in the name 
>>> of generality and transferability.
>>> 
 So basically the reason of not getting correct surface tension is bad 
 parametrisation of force field.
 Then I should not spend time on simulation the replica exchange.
>>> 
>>> Indeed.  Look at the things I mentioned before.  If the model doesn't work 
>>> well, then you have to get into the details of why.  Burning through more 
>>> time hoping it will work out and expensive stuff like REMD is just a waste.
>>> 
>>> -Justin
>>> 
>>> --
>>> ==
>>> 
>>> Justin A. Lemkul, Ph.D.
>>> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>>> 
>>> Department of Pharmaceutical Sciences
>>> School of Pharmacy
>>> Health Sciences Facility II, Room 629
>>> University of Maryland, Baltimore
>>> 20 Penn St.
>>> Baltimore, MD 21201
>>> 
>>> jalem...@outerbanks.umaryland.edu | (410) 706-7441
>>> http://mackerell.umaryland.edu/~jalemkul
>>> 
>>> ==
>>> --
>>> 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-requ...@gromacs.org.
>> 
> 
> -- 
> ==
> 
> Justin A. Lemkul, Ph.D.
> Ruth L. Kirschstein NRSA Postdoctoral Fellow
> 
> Department of Pharmaceutical Sciences
> School of Pharmacy
> Health Sciences Facility II, Room 629
> University of Maryland, Baltimore
> 20 Penn St.
> Baltimore, MD 21201
> 
> jalem...@outerbanks.umaryland.edu  
> | (410) 706-7441
> http://mackerell.umaryland.edu/~jalemkul 
> 
> 
> ==
> -- 
> 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-requ...@gromacs.org 
> .

-- 
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-requ...@gromacs.org.

Re: [gmx-users] Replica Exchange for surface tension

[Justin Lemkul]

On 3/29/17 9:16 AM, gozde ergin wrote:

Are you sure it is not CHARMM36 because it is written:

;;
;; Generated by CHARMM-GUI (http://www.charmm-gui.org) v1.7
;;
;; psf2itp.py
;;
;; Correspondance:
;; j712l...@ku.edu or won...@ku.edu
;;
;; CHARMM36 FF in GROMACS format
;;

on the top of the .itp files.



OK, I was thinking too quickly, the workflow goes like this: If a molecule is 
identified as part of the existing force field, it is used.  If it is not, then 
CGenFF is called.  So your SDS may be CHARMM36 but your cocobetaine is not.  You 
can check against the CHARMM force field files and also the atom types 
themselves.  If the second character is "G" then it is CGenFF, otherwise it is 
stock CHARMM.


Do not rely on header comments; those are often boilerplate.  Check the 
parameters themselves.


-Justin





On 29 Mar 2017, at 15:06, Justin Lemkul  wrote:



On 3/29/17 4:15 AM, gozde ergin wrote:

I used CHARMM GUI Input Generator tool both for SDS and cocobetaine.


For the sake of clarity, you should understand that you are *not* then using 
CHARMM36.  CHARMM-GUI calls the CGenFF program to parametrize the molecule, so 
you are using CGenFF parameters, which are compatible with CHARMM by design, 
but are not the standard CHARMM36 parameters.  Thus your SDS model will differ 
from the standard one that we have in the lipids/detergents file.  You should 
compare it to make sure it is reasonable.  A general force field will sacrifice 
some accuracy in the name of generality and transferability.


So basically the reason of not getting correct surface tension is bad 
parametrisation of force field.
Then I should not spend time on simulation the replica exchange.


Indeed.  Look at the things I mentioned before.  If the model doesn't work 
well, then you have to get into the details of why.  Burning through more time 
hoping it will work out and expensive stuff like REMD is just a waste.

-Justin

--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalem...@outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==
--
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-requ...@gromacs.org.




--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalem...@outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==
--
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-requ...@gromacs.org.

Re: [gmx-users] Replica Exchange for surface tension

[gozde ergin]

Are you sure it is not CHARMM36 because it is written:

;;
;; Generated by CHARMM-GUI (http://www.charmm-gui.org) v1.7
;;
;; psf2itp.py
;;
;; Correspondance:
;; j712l...@ku.edu or won...@ku.edu
;;
;; CHARMM36 FF in GROMACS format
;;

on the top of the .itp files.



> On 29 Mar 2017, at 15:06, Justin Lemkul  wrote:
> 
> 
> 
> On 3/29/17 4:15 AM, gozde ergin wrote:
>> I used CHARMM GUI Input Generator tool both for SDS and cocobetaine.
> 
> For the sake of clarity, you should understand that you are *not* then using 
> CHARMM36.  CHARMM-GUI calls the CGenFF program to parametrize the molecule, 
> so you are using CGenFF parameters, which are compatible with CHARMM by 
> design, but are not the standard CHARMM36 parameters.  Thus your SDS model 
> will differ from the standard one that we have in the lipids/detergents file. 
>  You should compare it to make sure it is reasonable.  A general force field 
> will sacrifice some accuracy in the name of generality and transferability.
> 
>> So basically the reason of not getting correct surface tension is bad 
>> parametrisation of force field.
>> Then I should not spend time on simulation the replica exchange.
> 
> Indeed.  Look at the things I mentioned before.  If the model doesn't work 
> well, then you have to get into the details of why.  Burning through more 
> time hoping it will work out and expensive stuff like REMD is just a waste.
> 
> -Justin
> 
> -- 
> ==
> 
> Justin A. Lemkul, Ph.D.
> Ruth L. Kirschstein NRSA Postdoctoral Fellow
> 
> Department of Pharmaceutical Sciences
> School of Pharmacy
> Health Sciences Facility II, Room 629
> University of Maryland, Baltimore
> 20 Penn St.
> Baltimore, MD 21201
> 
> jalem...@outerbanks.umaryland.edu | (410) 706-7441
> http://mackerell.umaryland.edu/~jalemkul
> 
> ==
> -- 
> 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-requ...@gromacs.org.

-- 
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-requ...@gromacs.org.

Re: [gmx-users] potential energy of a system

[Justin Lemkul]

On 3/29/17 8:43 AM, Mark Abraham wrote:

Hi,

The total potential energy has everything by definition, so I don't know
what you're asking.

Additivity of the components of the potential energy is an artefact of the
construction of the model physics. While it is true that different physical
phenomena underlie some of the components, the parameterization of the
force field did not target meaningful decomposition of the potential
energy. But you might be able to benefit from fortuitous cancellation
effects when comparing such things.



Nitpick here: in CHARMM (and our polarizable model), we seek to get a variety of 
QM interaction energies explicitly correct (interactions with water, ions, other 
small molecules) so indeed there is some reliability in these metrics with 
CHARMM.  Of course, like all force fields, there is some inherent error 
cancellation but we do make an effort to get all of this kind of stuff right. 
Most other force fields, of course, do not, seeking instead to get overall 
properties correct, leaving the details to be debatable.  We try to get it all :)


-Justin

--
==

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalem...@outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==
--
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-requ...@gromacs.org.

Re: [gmx-users] NVE run

[Neda Rafiee]

Thanks a lot Mark...

Kind regards,
Neda Rafieiolhosseini
Ph.D. Student at Institute for Research
in Fundamental Sciences(IPM),
School of Nano,
Tehran,Iran. 

- Original Message -
From: "Mark Abraham" 
To: gmx-us...@gromacs.org
Sent: Wednesday, March 29, 2017 2:38:54 PM
Subject: Re: [gmx-users] NVE run

Hi,

There's no way to say. Those numbers would be more indicative of a problem
if the system was small, or the simulation was short, but we don't know
those (and there are many algorithm settings that contribute). What you
really need is to know whether the quality of your observables is degraded
over the time scale of the simulation; at best energy conservation is only
a proxy for that. See thoughts at
http://biorxiv.org/content/early/2016/10/24/083055

Mark

On Wed, Mar 29, 2017 at 2:02 PM Neda Rafiee  wrote:

> Dear gromax users,
> I have already run my system under NVE condition with dt=2fs and I found
> the energy drift about -7000 KJ/mol, then I repeated my simulation with
> dt=1fs again and I found the drift about -1900 KJ/mol. Is it a reasonable
> value or should I have a smaller drift?  THANKS.
>
> --
> 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-requ...@gromacs.org.
>
-- 
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-requ...@gromacs.org.
-- 
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-requ...@gromacs.org.

Re: [gmx-users] importance of nvlink in GPU nodes?

[Mark Abraham]

Hi,

There's no use made of NVLINK in GROMACS. The 2018 release might make some
use of it.

Mark

On Wed, Mar 29, 2017 at 4:19 AM Christopher Neale <
chris.ne...@alum.utoronto.ca> wrote:

> Dear users:
>
> does anybody know if the presence of either CPU-to-GPU or GPU-to-GPU
> nvlink affects the performance or efficiency of gromacs?
>
> Thank you,
> Chris.
> --
> 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-requ...@gromacs.org.
>
-- 
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-requ...@gromacs.org.

Re: [gmx-users] potential energy of a system

[Mark Abraham]

Hi,

The total potential energy has everything by definition, so I don't know
what you're asking.

Additivity of the components of the potential energy is an artefact of the
construction of the model physics. While it is true that different physical
phenomena underlie some of the components, the parameterization of the
force field did not target meaningful decomposition of the potential
energy. But you might be able to benefit from fortuitous cancellation
effects when comparing such things.

Mark

On Wed, Mar 29, 2017 at 12:00 PM Saumyak Mukherjee <
mukherjee.saumya...@gmail.com> wrote:

> Dear users,
>
> Which terms are included in the total potential energy of a system?
> Does it have COUL-SR and LJ-SR only, or does it also include the 1-4 terms?
>
> Thanks in advance.
>
> Regards,
> Saumyak
>
> --
> 
> *Saumyak Mukherjee*
>
> Junior Research Fellow
> Prof. Biman Bagchi's Group
> Solid State and Structural Chemistry Unit
> Indian Institute of Science
> Bangalore - 560012
>
> Mob : 8017292426
> Alternative e-mail : saumyakmukher...@gmail.com
> smukher...@sscu.iisc.ernet.in
> 
> --
> 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-requ...@gromacs.org.
>
-- 
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-requ...@gromacs.org.

Re: [gmx-users] NVE run

[Mark Abraham]

Hi,

There's no way to say. Those numbers would be more indicative of a problem
if the system was small, or the simulation was short, but we don't know
those (and there are many algorithm settings that contribute). What you
really need is to know whether the quality of your observables is degraded
over the time scale of the simulation; at best energy conservation is only
a proxy for that. See thoughts at
http://biorxiv.org/content/early/2016/10/24/083055

Mark

On Wed, Mar 29, 2017 at 2:02 PM Neda Rafiee  wrote:

> Dear gromax users,
> I have already run my system under NVE condition with dt=2fs and I found
> the energy drift about -7000 KJ/mol, then I repeated my simulation with
> dt=1fs again and I found the drift about -1900 KJ/mol. Is it a reasonable
> value or should I have a smaller drift?  THANKS.
>
> --
> 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-requ...@gromacs.org.
>
-- 
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-requ...@gromacs.org.

[gmx-users] NVE run

[Neda Rafiee]

Dear gromax users,
I have already run my system under NVE condition with dt=2fs and I found the 
energy drift about -7000 KJ/mol, then I repeated my simulation with dt=1fs 
again and I found the drift about -1900 KJ/mol. Is it a reasonable value or 
should I have a smaller drift?  THANKS.

-- 
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-requ...@gromacs.org.

[gmx-users] potential energy of a system

[Saumyak Mukherjee]

Dear users,

Which terms are included in the total potential energy of a system?
Does it have COUL-SR and LJ-SR only, or does it also include the 1-4 terms?

Thanks in advance.

Regards,
Saumyak

-- 

*Saumyak Mukherjee*

Junior Research Fellow
Prof. Biman Bagchi's Group
Solid State and Structural Chemistry Unit
Indian Institute of Science
Bangalore - 560012

Mob : 8017292426
Alternative e-mail : saumyakmukher...@gmail.com
smukher...@sscu.iisc.ernet.in

-- 
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-requ...@gromacs.org.

Re: [gmx-users] Replica Exchange for surface tension

[gozde ergin]

I used CHARMM GUI Input Generator tool both for SDS and cocobetaine. 
So basically the reason of not getting correct surface tension is bad 
parametrisation of force field.
Then I should not spend time on simulation the replica exchange.

Thanks Justin.

> On 28 Mar 2017, at 22:48, Justin Lemkul  wrote:
> 
> 
> 
> On 3/28/17 2:59 PM, gozde ergin wrote:
>> I use Charmm36 force field. Do you mean the force field parameter between
>> cocobetaine and SDS? Because for pure surfactant I am able to estimate the
>> correct surface tension. Which means force field parameter of SDS-water and
>> cocobetaine-water is fine.
>> 
> 
> SDS is a standard part of CHARMM36 but I have no idea what cocobetaine is.  
> Did you find existing parameters for it or did you do the parametrization 
> yourself? The larger point is that you've probably stumbled upon the corner 
> cases that often come up - some combination of parameters has never been 
> looked at before and it requires more refinement.  If both interact with 
> water fine but interact with each other improperly, that means either the 
> SDS-cocobetaine interactions are too weak or too strong, or that they induce 
> some sort of abnormal conformational sampling.
> 
> You'll need to look at the microscopic details of how they interact - which 
> groups are associated with which, how the conformational ensembles of each 
> differ from those observed in water, etc.  Then you can focus on specific 
> elements of refinement that may need to be done (e.g. NBFIX for LJ 
> interactions that are often the culprit in such cases, and/or perhaps some 
> sub-par dihedral parameters that lead to bad sampling when a particular 
> interaction occurs).
> 
> -Justin
> 
>> Sent from my iPhone
>> 
>>> On 28 Mar 2017, at 7:21 PM, Justin Lemkul  wrote:
>>> 
>>> 
>>> 
 On 3/28/17 12:43 PM, gozde ergin wrote: I have a mixed system of SDS and
 cocobetaine on water surface. I estimated the surface tension of SDS on
 water surface and cocabetaine on surface tension separately and correctly
 (comparing with experimental results). However I could not estimate the
 correct surface tension for SDS-cocobetaine mix in even 150 ns
 simulation. Do you think replica exchange simulation would help me to
 estimate the correct surface tension of mixed system? Do you have any
 idea why I failed in estimating the mixed system surface tension?
 
>>> 
>>> If a physical model fails to produce known physical properties, the problem
>>> is often the model itself.  I'd suspect the force field parameters first.
>>> What is the source of your parameters?
>>> 
>>> Replica exchange helps you get over large barriers to improve sampling, but
>>> I doubt that's useful here in a mixture of small molecules.
>>> 
>>> -Justin
>>> 
>>> -- ==
>>> 
>>> Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow
>>> 
>>> Department of Pharmaceutical Sciences School of Pharmacy Health Sciences
>>> Facility II, Room 629 University of Maryland, Baltimore 20 Penn St.
>>> Baltimore, MD 21201
>>> 
>>> jalem...@outerbanks.umaryland.edu | (410) 706-7441
>>> http://mackerell.umaryland.edu/~jalemkul
>>> 
>>> == -- 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-requ...@gromacs.org.
> 
> -- 
> ==
> 
> Justin A. Lemkul, Ph.D.
> Ruth L. Kirschstein NRSA Postdoctoral Fellow
> 
> Department of Pharmaceutical Sciences
> School of Pharmacy
> Health Sciences Facility II, Room 629
> University of Maryland, Baltimore
> 20 Penn St.
> Baltimore, MD 21201
> 
> jalem...@outerbanks.umaryland.edu  
> | (410) 706-7441
> http://mackerell.umaryland.edu/~jalemkul 
> 
> 
> ==
> -- 
> 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-requ...@gromacs.org 
> .

-- 
Gromacs Users mailing list

* Please search the archive at