Hi Tinashe,
Am 31. Januar 2012 08:12 schrieb Tinashe <[email protected]>:
> I have looked at the tutorial handouts and all the parts of the manual
> and many forum discussions.
>
> This is what I am understanding so far (step 1 and 2 are complete):
> 1. To map from the atomistic to coarse-grained:
> csg_map --top topol.tpr --trj confout.gro --cg propane.xml --out
> conf_cg.gro
> 2. To make the coarse grained topology file:
> csg_gmxtopol --top topol.tpr --cg propane.xml --out out.top
out.top is just a template, you will still need to check it manually
for correctness.
> 3. To calculate distributions:
> csg_stat --top topol.tpr --trj traj.trr --cg propane.xml --nt 3 --
> options fmatch.xml
> Output files: CH2-CH3.dist.new, CH2-CH3.dist.new ,CH2-
> CH3.dist.new, bond.dist.new, angle.dist.new
> Some of these files will be renamed in the IBI CG run to:
> CH2-CH3.dist.tgt, CH2-CH3.dist.tgt, CH2-CH3.dist.tgt
> What use are these files then?: bond.dist.new, angle.dist.new
You could boltzmann invert bond.dist.new, angle.dist.new by hand using:
$ awk -v kbt=NUMBER '{print $1,($1>0)?-kbt*log($2/$1/$1):"nan"}'
bond.dist.new > bond.pot.new
However csg_boltzmann takes care of the normalization (above /$1/$1)!
> 4. To calculate the potentials (I do not want to do it manually):
You only have to do it one time. If you know a smart way to do it
automatically tell me.
Votca 1.3 supports IBI of bonded interaction now, but you still need
to give an initial guess for the potential.
> csg_boltzmann --top topol.tpr --trj traj.trr --cg propane.xml
> Output files: angle.dist.ib, angle.pot.ib, bond.dist.ib,
> bond.pot.ib
> (a) manually delete poorly sampled parts of: "bond.pot.ib" to
> "bond.pot.ib_new"
> and "angle.pot.ib" to "angle.pot.ib_new"
> (b) Execute the following commands to convert to the formatting of
> Gromacs:
> csg_call --options bond.xml --ia-type bonded
> convert_potential gromacs bond.pot.ib_new table_b1.xvg
> csg_call --options angle.xml --ia-type bonded
> convert_potential gromacs bond.pot.ib_new table_a1.xvg
> The shape of these 2 graphs are similar to what you have in
> the tutorial but the numbers are different??
At some point we decides to shift them, which is that difference.
However constants in the potential do not not matter for the physics.
> 5. Coarse-grained run:
> (a) Copy table.xvg from an example
> (b) Copy table_b1.xvg and table_a1.xvg from step 4 as well as the
> CG coordinates and topology files
> (c) Execute the following command which will run the CG simulation
> and produce new files and tables:
> (e.g. table_CH2-CH3.xvg, table_CH2-CH3.xvg, table_CH2-CH3.xvg
> as specified in the settings file)
> csg_inverse --options settings.xml
No, (*.xvg or *.dist.tgt) are the targets, which are inputs.
csg_inverse will generate *.pot.new, which are the IBI potentials.
It also generates *.dist.new, which are the distributions of the IBI
simulations.
Once you think, you have all input files, run:
$ grompp -n index.ndx
$ mdrun -v
to see if you have missed an input file.
Christoph
> Parts of step 3 - 5 are not correct in my understanding for sure
> otherwise everything will be fine and this is what we would discuss.
> Perhaps others on the forum can clarify where I am missing the linking
> in the above 5 steps.
>
> Thanks again for your time and help.
>
> Tinashe
>
> On Jan 21, 10:44 pm, Christoph Junghans <[email protected]> wrote:
>> Am 20. Januar 2012 09:41 schrieb Tinashe <[email protected]>:> Before
>> I do what you have suggested, I do not know yet 100% how to get
>> > both bonded and non-bonded potentials in VOTCA "correctly".
>>
>> > 1. For instance, how do you obtain the table.pot file?
>> > This file is needed to start the following sequence of commands:
>>
>> I guess, you are speaking about those non-bonded interaction which are
>> not iterated, usually named table.xvg and table_X_Y.xvg in the
>> maindir.
>>
>> table.xvg is Gromacs standard interaction table for all non-bonded
>> interaction, which are not covered by the pairs listed in
>> energygrp_table of your mdp file.
>> In all examples of csg-tutorial this file has no meaning as all
>> interaction are covered by energygrp_table, so you can fill it with
>> zeros. However gromacs still needs it. (see Gromacs Manual for
>> details)
>>
>> table_X_Y.xvg, where X and Y are groups from energygrps of your mdp file,
>> can come for a former ibi simulation or somewhere else. These are only
>> needed if you do a iteration of only some of the non-bonded
>> interactions, no csg tutorial does that.
>>
>>
>>
>> > csg_sample --in table.pot --out table_resample.pot --grid min:step:max
>> > csg_call table extrapolate table_resample.pot table_extrapolate.pot
>> > csg_call --options table.xml convert_potential gromacs
>> > table_extrapolate.pot table.xvg
>>
>> > 2. Similarly, how do I obtain the table.pot for both the bonded and
>> > angle potentials
>> > which I will resample, extrapolate and later convert to table_b1.xvg
>> > and table_a1.xvg?
>>
>> These are bonded interactions. See chapter 5 of the manual on how to
>> calculate them. Basically you do simple Boltzmann inversion using
>> csg_boltzmann or a simple script.
>>
>>
>>
>> > Is the way of obtaining the target non-bonded potentials for the IBI
>> > method through force matching (csg_fmatch)
>> > or are the potentials only supposed to be derived using csg_stat and/
>> > or csg_boltzmann?
>>
>> IBI aims to reproduce distributions (structure based coarse-graining), while
>> force matching aims to reproduce force.
>>
>> csg_stat only calculates bonded and non-bonded distributions.
>> csg_boltzmann calculates bonded distribution and other analysis
>> including Boltzmann inversion.
>>
>> Christoph
>>
>>
>>
>>
>>
>>
>>
>> > Perhaps a step-by-step explanation will help me to reproduce your
>> > hexane tutorial and serve to help the other new VOTCA users as well.
>>
>> > Thanks a lot.
>>
>> > Tinashe
>>
>> > On Jan 18, 9:36 pm, Christoph Junghans <[email protected]> wrote:
>> >> Hi Tinashe,
>>
>> >> welcome!
>>
>> >> Am 18. Januar 2012 07:18 schrieb Tinashe <[email protected]>:> -
>> >> the min/max/step values for force matching i get from the
>> >> > distribution functions, that it samples a valid interval, and the
>> >> > number of bins is not too high, otherwise might become slow
>> >> > - the table_a1*.xvg are usually generated using csg_boltzmann, there
>> >> > is a section in the manual. You can convert the votca potential files
>> >> > to gromacs using csg_call. I have to lookup the correct command, but
>> >> > christoph might answer on googlrgroups.
>>
>> >> See section 4.2 of the manual:
>> >> <https://votca.googlecode.com/files/votca-csg-manual-1.2.2.pdf>
>>
>> >> But basically once you have the potential in Votca format, it boils
>> >> down to the following:
>> >> -create a simple settings xml file (table.xml):
>> >> <cg>
>> >> <inverse>
>> >> <gromacs>
>> >> <pot_max>1e8</pot_max>
>> >> <table_end>8.0</table_end>
>> >> <table_bins>0.002</table_bins>
>> >> </gromacs>
>> >> </inverse>
>> >> </cg>
>> >> and run
>> >> $csg_call --ia-type TYPE --options table.xml convert_potential gromacs \
>> >> table_extrapolate.pot table.xvg
>> >> where type can be non-bonded, bonded, angle, dihedral.
>>
>> >> Note, that in the current development version we renamed "bonded" to
>> >> "bond"!
>>
>> >> Cheers,
>>
>> >> Christoph
>>
>> >> > - renaming te distribution function and use it as target is the proper
>> >> > way to do
>> >> > - you can make a few optimizations in the grompp, e.g. if you don't
>> >> > have charge in the cg model, use cut-off as estatic method. this will
>> >> > disable the fft from pme.
>>
>> >> > Best,
>> >> > Victor
>>
>> >> >> Dear Victor,
>>
>> >> >> Firstly, thanks a lot, your suggestions were on the mark.
>>
>> >> >> At first, I ran the following command:
>> >> >> csg_stat --top topol.tpr --trj traj.trr --cg hexane.xml --nt 3
>> >> >> --options
>> >> >> fmatch.xml
>> >> >> and then renamed the files from *.dist.new to *.dist.tgt to use
>> >> >> together
>> >> >> with the settings file. Is this the correct way to things?
>>
>> >> >> How are the following files meant to be created: table.xvg table_a1.xvg
>> >> >> table_b1.xvg
>> >> >> Where does one get the numbers in the settings.xml and fmatch.xml? I
>> >> >> have
>> >> >> simply used numbers from VOTCA tutorials?
>> >> >> It appears that I can still use the same grommp file as the one I used
>> >> >> for
>> >> >> atomistic simulations, is this so?
>>
>> >> >> Kind regards,
>> >> >> Tinashe
>>
>> >> > --
>> >> > You received this message because you are subscribed to the Google
>> >> > Groups "votca" group.
>> >> > To post to this group, send email to [email protected].
>> >> > To unsubscribe from this group, send email to
>> >> > [email protected].
>> >> > For more options, visit this group
>> >> > athttp://groups.google.com/group/votca?hl=en.
>>
>> >> --
>> >> Christoph Junghans
>> >> Web:http://www.compphys.de-Hide quoted text -
>>
>> >> - Show quoted text -
>>
>> > --
>> > You received this message because you are subscribed to the Google Groups
>> > "votca" group.
>> > To post to this group, send email to [email protected].
>> > To unsubscribe from this group, send email to
>> > [email protected].
>> > For more options, visit this group
>> > athttp://groups.google.com/group/votca?hl=en.
>>
>> --
>> Christoph Junghans
>> Web:http://www.compphys.de- Hide quoted text -
>>
>> - Show quoted text -
>
> --
> You received this message because you are subscribed to the Google Groups
> "votca" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected].
> For more options, visit this group at
> http://groups.google.com/group/votca?hl=en.
>
--
Christoph Junghans
Web: http://www.compphys.de
--
You received this message because you are subscribed to the Google Groups
"votca" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/votca?hl=en.
