generate tabulated potential for non-bonded interaction

[Zhicheng]

Dear all:

The "hist" ("tab") command within csg_boltzmann tools to generate the distribution function (tabulated potential), after post-processing, can be used to derive bonded interaction ( e.g *.dist.tgt table_a1.tab table_b1.tab in csg-tutorials/propane/ibi_espresso). However, the tabulated potential required by non-bonded interaction can't obtain using this method. Therefore, I want to know how to generate the tabulated potential for non-bonded interaction ( such as table_CG_CG.tab in csg-tutorials/methanol/ibi_espresso/generate_esp_from_gro). I will be appreciative if anyone can help me!

Best Regards!

Zhicheng

[Christoph Junghans]

2013/6/28 Zhicheng <zcgu...@gmail.com>:

I think you mixing up two thing here.

1.) csg_inverse will generate table_CG_CG.tab during IBI procedure
for you, the initial potential is calculated out of the rdf
(rdf_methanol.xvg). So for IBI there is no initial table_CG_CG.tab
needed.
2.) The files in generate_esp_from_gro are converts from gromacs input
to espresso input (see generate.tcl). This was mainly done to allow
user to compare the behavior of different simulation back-ends
(espresso vs gromacs). In a real world study, there are very few cases
where one would convert gromacs input files to espresso instead of
simply using gromacs directly.

Christoph
>
> Best Regards!
>
> Zhicheng
>
> --
> You received this message because you are subscribed to the Google Groups
> "votca" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to votca+un...@googlegroups.com.
> To post to this group, send email to vo...@googlegroups.com.
> Visit this group at http://groups.google.com/group/votca.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>



--
Christoph Junghans
Web: http://www.compphys.de

[Zhicheng]

Dear Christoph:
Thanks for your reply. It is very useful for me!
在 2013年6月28日星期五UTC+8下午10时49分24秒，Christoph Junghans写道：

[Zhicheng]

Hi Christoph,

I retried the methanol/ibi_espresso tutorial and modified the bead interaction of lennard-jones (inter 0 0 lennard-jones 1.0 1.0 1.12246 0.25 0.0 ) to replace the tabulated potential (inter 0 0 tabulated "table_CG_CG.tab" ) in generate.tcl. However, I got the following error message:

                                                                                   property not found: table

In my opinion, it means that the IBI procedure interfaced with espresso only requires tabulated potential other than another type of interaction (such as lennard-jones). Thus, we have to specify the initial guess of tabulated potential (table_CG_CG.tab). Is that right?

Best Regards!

Zhicheng
在 2013年6月28日星期五UTC+8下午10时49分24秒，Christoph Junghans写道：

[Christoph Junghans]

2013/7/3 Zhicheng <zcgu...@gmail.com>:

> Hi Christoph,
>
> I retried the methanol/ibi_espresso tutorial and modified the bead
> interaction of lennard-jones (inter 0 0 lennard-jones 1.0 1.0 1.12246 0.25
> 0.0 ) to replace the tabulated potential (inter 0 0 tabulated
> "table_CG_CG.tab" ) in generate.tcl. However, I got the following error
> message:
>
>
> property not found: table

Like I said before the scripts in generate_esp_from_gro are only to
generate espresso input data from gromacs input data.
(It basically generates conf.esp.gz from conf.gro)
If you have gromacs input data already just use gromacs to do IBI, it
is easier and quicker!

>
> In my opinion, it means that the IBI procedure interfaced with espresso only
> requires tabulated potential other than another type of interaction (such as
> lennard-jones). Thus, we have to specify the initial guess of tabulated
> potential (table_CG_CG.tab). Is that right?

No, table_CG_CG.tab is generated from the initial rdf.

To run IBI using Espresso just execute run.sh in methanol/ibi_espresso.
You might need to modify settings.xml if your Espresso binary is named
differently.
run.sh execute csg_inverse, which then calls Espresso.

Remember to use the updated conf.esp.gz from:
<https://csg-tutorials.votca.googlecode.com/hg/methanol/ibi_espresso/conf.esp.gz>

Christoph

[郭志成]

 Hi Christoph,

Thanks for your reply. Hmm, you're right that it is more quicker and easier to run IBI using gromacs than espresso. As mentioned above, the initial tabulated potential for bonded interacation directly can be obtained using csg_boltzmann tool, however, initial guess of tabulated potential for non-bonded interaction can not be generated by csg_boltzamnn tool. Thus, I want to know how to get the initial non-bonded tabulated potential (csg-tutorial/methanol/ibi/table.xvg) through calculating rdf in VOTCA.

Sincerely,

Zhicheng

2013/7/3 Christoph Junghans <jung...@votca.org>

You received this message because you are subscribed to a topic in the Google Groups "votca" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/votca/1ab9M7DmThk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to votca+un...@googlegroups.com.

[Christoph Junghans]

2013/7/4 郭志成 <zcgu...@gmail.com>:

> Hi Christoph,
> Thanks for your reply. Hmm, you're right that it is more quicker and easier
> to run IBI using gromacs than espresso. As mentioned above, the initial
> tabulated potential for bonded interacation directly can be obtained using
> csg_boltzmann tool, however, initial guess of tabulated potential for
> non-bonded interaction can not be generated by csg_boltzamnn tool. Thus, I
> want to know how to get the initial non-bonded tabulated potential
> (csg-tutorial/methanol/ibi/table.xvg) through calculating rdf in VOTCA.

You don't need to provide an initial guess. VOTCA is doing that
automatically, you could do the same manually using:
$ csg_call --ia-name CG-CG --options settings.xml --ia-type non-bonded
prepare_single ibi
(change the "CG-CG" to name of your interaction)
This will generate a CG-CG.pot.new, which is the initial guess
calculated from the rdf. You could preprocess that file and give it to
Votca as initial guess (CG-CG.pot.in).

Please don't confuse table.xvg with an initial guess. The interaction
between the CG beads in methanol/ibi is given as table_CG_CG.xvg to
gromacs (see the xml settings file). table.xvg is used for all other
interactions, which are not covered by special tables (table_X_Y.xvg).
Even if there are no other interactions gromacs will need a table.xvg
file (this might has changed in gromacs 4.6). None of our tutorials
actually uses table.xvg, this is just a gromacs specialty.

Christoph

[郭志成]

Dear Christoph,
Thanks for your patient interpretation. I had a misunderstanding about initial guess of non-bonded tabulated potential and other interactions (table.xvg)

2013/7/5 Christoph Junghans <jung...@votca.org>

[郭志成]

And, how can I generate other intercations table.xvg for gromacs.

 

Sincerely,

Zhicheng

2013/7/5 郭志成 <zcgu...@gmail.com>

[Christoph Junghans]

2013/7/4 郭志成 <zcgu...@gmail.com>:

> And, how can I generate other intercations table.xvg for gromacs.

I don't really understand your question!

For the interactions, which you iterate on an no initial guess is
need, you just need the rdf.
For all others (fixed) interaction, you will need a table. Manual
describes this procedure for bonded interaction in some details, for
non-bonded interactions use
CG-CG.pot.new from

$ csg_call --ia-name CG-CG --options settings.xml --ia-type non-bonded
prepare_single ibi

together with
$ csg_call convert_potential gromacs CG-CG.pot.new table_CG_CG.xvg
to convert CG-CG.pot.new to table_CG_CG.xvg.

[Zhicheng]

Dear Christoph,

I apologize to you for my question. I have known how to get the tabulated potential for bonded and non-bonded interactions.The table.xvg in methanol/ibi you metioned is the file required by gromacs even if there are no other interactions. My question is that how to generate this file. Whether is it similar to bonded or non-bonded interaction. In addition, what is the type of interaction represented by table.xvg because that the interaction between the CG beads in methanol/ibi is given as table_CG_CG.xvg.

 

Sincerely,

Zhicheng

在 2013年7月6日星期六UTC+8上午12时01分18秒，Christoph Junghans写道：

[Christoph Junghans]

2013/7/6 Zhicheng <zcgu...@gmail.com>:

> Dear Christoph,
> I apologize to you for my question. I have known how to get the tabulated
> potential for bonded and non-bonded interactions.The table.xvg in
> methanol/ibi you metioned is the file required by gromacs even if there are
> no other interactions. My question is that how to generate this file.

table.xvg isn't used anywhere, so you could use a file with zeros
(like it is done in spce/ibi and the other examples), which I think is
the safest.
Where the table.xvg in methanol/ibi cames from, I don't remember, it
seems to be simple BI potential.

> Whether is it similar to bonded or non-bonded interaction.

Non-bonded tables have 7 columns while bonded tables have 3 columns.
Bonded tables usually come from a Boltzmann inversion of the bonded
distribution.

> In addition, what
> is the type of interaction represented by table.xvg because that the
> interaction between the CG beads in methanol/ibi is given as
> table_CG_CG.xvg.

Like I said before table.xvg isn't used in this simulation, but
gromacs uses it for the interaction of the "rest" energy group.
Gromacs require table.xvg, even if the rest group is empty.

In the case methanol/ibi the rest group is empty as the whole system
belongs to the CG energy group (see index.ndx) and the CG-CG
interaction is covered by a special interaction (grompp.mdp:
energygrp_table = CG CG).

table.xvg could be useful for a multi-component system, where only one
interactions is iterated and all the others are covered normal LJ
interactions.

Christoph

[郭志成]

Dear Christoph,

Thanks for your help. It is working now. There is another question about the coarse-grained topology file as shown in methanol/ibi/topol.top and propane/ibi/topol.top. I have been familiar with the atomistic topology. Therefore, I want to know how to determine these parameters related to coarse-grained beads (e.g. c6, c12, bonds and angles items).Whether the specification of these parameters are arbitrary.

 

Sincerely,

Zhicheng

2013/7/8 Christoph Junghans <jung...@votca.org>

[Christoph Junghans]

2013/7/7 郭志成 <zcgu...@gmail.com>:
> Dear Christoph,
>

> Thanks for your help. It is working now. There is another question about the
> coarse-grained topology file as shown in methanol/ibi/topol.top and
> propane/ibi/topol.top. I have been familiar with the atomistic topology.
> Therefore, I want to know how to determine these parameters related to
> coarse-grained beads (e.g. c6, c12, bonds and angles items).Whether the
> specification of these parameters are arbitrary.

No, it is not arbitrary, if you are using tables in gromacs the values
c6 and c12 are the pre-factors for the table column 4/5 (c6) and 6/7
(c12). (also see gromacs manual section 4.2.13)
In VOTCA we only use column 6 and 7 for IBI hence we set c12 to 1.0
(and c6 to 0, which not 100% necessary as column 4,5 are filled with
zeros).

In some cases (e.g. spce/tf) it is useful to have c6 and c12 set to
something non-trivial.

Christoph

[Zhicheng]

Dear Christoph,
I can't get the parameters by fitting related properties (thermodynamics, hydrodynamics) because there are no experimental and theoretical results about the coarse-grained beads. Thus, how can I obtain the bonded prameters in topology file (e.g. bonds and angles items in propane/ibi/topol.top)
在 2013年7月8日星期一UTC+8下午12时45分52秒，Christoph Junghans写道：

[Christoph Junghans]

2013/7/8 Zhicheng <zcgu...@gmail.com>:
> Dear Christoph,

> I can't get the parameters by fitting related properties (thermodynamics,
> hydrodynamics) because there are no experimental and theoretical results
> about the coarse-grained beads. Thus, how can I obtain the bonded prameters
> in topology file (e.g. bonds and angles items in propane/ibi/topol.top)

You mean the following part, right?
[ bonds ]
; ai aj funct c0 c1 c2 c3
1 2 8 1 1.0
2 3 8 1 1.0

[ angles ]
; ai aj ak funct c0 c1 c2 c3
1 2 3 8 1 1.0
funct=8 means use a table with number c0=1 and scale it with c1=1.0
(see gromacs manual for details)

The used tables table_a1.xvg (a for angle) and table_b1.xvg (b for
bond) are obtained from a Boltzmann inversion of the atomistic
bond/angle distribution. (see <http://dx.doi.org/10.1021/ct900369w>
for details)