Re: Re: vmd-l: making psf files

From: Kenno Vanommeslaeghe (kvanomme_at_rx.umaryland.edu)
Date: Mon Oct 13 2014 - 12:18:31 CDT

- Please let me know what part of the warning on top of the page
http://mackerell.umaryland.edu/~kenno/tutorial/ is not clear.

- The warning contains a link to the download page
http://mackerell.umaryland.edu/~kenno/cgenff/download.html , which also
contains a note in Italic. Please let me know what part of it is unclear.
As you can see, you're using a 5-years-old alpha version on CGenFF; 6
newer (beta) versions have been released since then.

- Assigning atom types (and an initial guess) manually on non-biomolecules
is generally not recommended because the atom typing scheme has become too
complex to reliably apply in wetware. I'm using the CGenFF program at
paramchem.org for this purpose when I start working on a new molecule; so
should you.
http://mackerell.umaryland.edu/~kenno/cgenff/program.html

- In my opinion, the most important use of ffTK is to refine parameters
and charges that are guessed badly by the CGenFF program (the penalties
are a good indication for that). Ethane is a relative poor test system for
that purpose, because it is already in the CGenFF main file (RESI ETHA)
and doesn't require new parameters. Also, it's a bit too trivial in my
opinion. The *new* tutorial (available from the above download page) has a
fully worked-out parametrization example for the drug-like molecule
acetazolamide. As a training exercise, something like this would be a good
choice to try to reproduce using ffTK. Or, if that's too complex,
something like the central dihedral in 1-aminomethoxymethane may be a
better choice (a lot of simple hemiacetals, hemiaminals and related
compounds have not been parameterized because they're unstable).

- I'm not sure autopsf supports non-biomolecules. You'd probably have more
success calling psfgen directly.

- I hope you have a senior person in your lab who is experienced in this
field, else it will be very difficult to get anywhere.

- More resources:
        * http://mackerell.umaryland.edu/~kenno/cgenff/faq.html
        * https://cgenff.paramchem.org/commonFiles/faq.php
        * Search the CHARMM "parameter set discussion" forum

On 10/13/2014 05:55 AM, Zeinab Emami wrote:
> Dear Ashar Malik,
>
> Thank you fro your answer, this is my pdb file:
>
> *HETATM 1 C2 C 1 0.756 -0.045 -0.137 1.00 0.00 *
> *HETATM 2 C1 C 1 -0.757 0.044 0.138 1.00 0.00 *
> *HETATM 3 H3 C 1 -1.280 0.345 -0.782 1.00 0.00 *
> *HETATM 4 H4 C 1 -0.943 0.788 0.926 1.00 0.00 *
> *HETATM 5 H5 C 1 -1.127 -0.938 0.465 1.00 0.00 *
> *HETATM 6 H6 C 1 1.105 -1.069 0.059 1.00 0.00 *
> *HETATM 7 H7 C 1 1.290 0.657 0.521 1.00 0.00 *
> *HETATM 8 H8 C 1 0.953 0.215 -1.188 1.00 0.00 *
>
>
> ETHANE MOLECULE, [file also attached to the email].
>
> I used autopsf for the pdb file. I took the atom names form the charmm
> force field atom name list:
>
> http://dogmans.umaryland.edu/~kenno/tutorial/par_all36_cgenff.prm
> <http://dogmans.umaryland.edu/%7Ekenno/tutorial/par_all36_cgenff.prm>
>
> I got ethane as a sample for training in order to leran how to prepare the
> parameter fiel for any given new molecule.
> I am using fftk for this purpose and as soon as I have the pdb file, I
> shall be able to produce the parameter file for the
> given molecule.
>
> I know many people are doing this at the moment, though I did not managed
> to find a good and thorough relevant
> document on the web. I would appreciate it if you guide me to a beneficial
> document about this.
>
> Sincerely,
> Zeinab
>
>
>
>
>
>
> On Mon, Oct 13, 2014 at 12:40 PM, Ashar Malik <asharjm_at_gmail.com
> <mailto:asharjm_at_gmail.com>> wrote:
>
> Hi,
>
> Could you send me the PDB file that you are using to generate the PSF
> file. I will try looking at it and if I can let you know what is
> causing the problem.
>
> Also mention what force field are you using to generate the PSF file.
>
> The more information you provide - the quicker you will get your answer.
>
> Best,
> /A
>
> On Mon, Oct 13, 2014 at 10:00 PM, Zeinab Emami <zemami_at_ku.edu.tr
> <mailto:zemami_at_ku.edu.tr>> wrote:
>
> Dear Christopher Mayne,
>
> First, I like to thank you again for your informative answer to my
> question about producing psf file when getting the error of "Need
> to have at least one atom in your selection!".
>
> I tried your guideline and changed the names, types and charges of
> each atom in the pdb file. Though
>
> - when I change every parameter [atom name, atom type, atom
> charge], NO CHANGE AT ALL is visible in the output pdb file but in
> the atom names.
>
> - Although the atom names change, the problem resists and I get
> the same error when I use autopsf.
>
> I really don't expect you to tell me the last answer. I tried to
> search for a solution but failed. If you guide me to
> a useful document about this I would be grateful.
>
> Sincerely,
> Zeinab
>
> On Thu, Oct 2, 2014 at 7:03 PM, Mayne, Christopher G
> <cmayne2_at_illinois.edu> wrote:
>
> Zeinab,
>
> Please cc VMD-L so that others can learn from the discussion.
>
> Hand editing PDB or PSF files is generally a bad idea. These
> formats are technically adhere to fixed-width specifications
> such that minor mistakes can easily result in broken files.
> My workflow in VMD typically looks like this:
>
> 1) load the PDB file into VMD
>
> 2) use VMD's atomselection mechanisms to modify atom name,
> type, charge, etc for each atom
> e.g.,
> set sel [atomselect top "index 0"]
> $sel set name C1
> $sel set type CG2R61
> $sel set charge -0.115
> $sel delete
>
> 3) check that VMD has correctly identified bonds. If it has
> made a mistake, use the "topo addbond <id1> <id2>" or "topo
> delbond <id1> <id2>" to fix the connectivity
>
> 4) once bonds are correctly defined, angles and dihedrals can
> automatically be determined via
> topo guessangles
> topo guessdihedrals
>
> 5) If you have any impropers, it is best to add those
> manually, although topotools can make a guess
> topo addimproper <id1> <id2> <id3> <id4>
>
> 6) make sure you set the resname, chain, and segname as desired
>
> 7) tell VMD to update the topology
> mol reanalyze top
>
> 8) write the psf/pdb
> [atomselect top all] writepsf mymol.psf
> [atomselect top all] writepdb mymol.pdb
>
>
> typing "topo" with no arguments in VMD's tkcon will print an
> extensive usage message
>
> Regards,
> Christopher Mayne
>
> On Oct 2, 2014, at 5:42 AM, Zeinab Emami wrote:
>
>> Dear Christopher Mayne,
>>
>> I have the same problem. I tried to follow you advise,
>> though could not figure out how it wors up to now. The
>> tutorial seems not comprehensive.
>>
>> I also read this page:
>> http://www.ks.uiuc.edu/Research/vmd/plugins/topotools/ which
>> did not help much as well.
>>
>> Would you please give me a short hint on *_the procedure for
>> producing PSF file from a given pdb file that is void of
>> parameter info._*
>>
>> E.G. I tried to modify this pdb file for ethane, in order to
>> get its psf file:
>>
>> HETATM 1 C 0 -0.772 0.000 -0.000
>> C
>> HETATM 2 C 0 0.772 0.000 -0.000
>> C
>> HETATM 3 H 0 1.179 1.012 -0.083
>> H
>> HETATM 4 H 0 1.178 -0.578 -0.835
>> H
>> HETATM 5 H 0 1.178 -0.434 0.918
>> H
>> HETATM 6 H 0 -1.178 -0.432 0.919
>> H
>> HETATM 7 H 0 -1.178 -0.580 -0.834
>> H
>> HETATM 8 H 0 -1.179 1.012 -0.086
>> H
>> END
>>
>> I changed the atom names and type manually, and the charges
>> as well, it did not work.
>>
>> I thank you in advance and look forward to hear your advise.
>>
>> Sincerely,
>> Zeinab
>>
>> On Wed, Oct 1, 2014 at 5:51 PM, Mayne, Christopher G
>> <cmayne2_at_illinois.edu> wrote:
>>
>> MK,
>>
>> If you have the mol2 and/or VMD's automated bond
>> detection is correct, you can use the TopoTools plugin
>> (https://sites.google.com/site/akohlmey/software/topotools) all
>> of the bonded elements found in a PSF (e.g., bonds,
>> angles, dihedrals). TopoTools has a function for
>> guessing impropers, but it is better to set those
>> manually where required. I frequently use this
>> technique to construct PSF files for small molecule
>> ligands from PDB files. If not included in your input,
>> you will also need to manually set the atom types and
>> partial charges using VMD's atom selection language.
>>
>> Regards,
>> Christopher Mayne
>>
>> On Oct 1, 2014, at 8:02 AM, John Stone wrote:
>>
>>> Hi,
>>> If you have a mol2 file with bond information, you may
>>> not
>>> really even need a PSF file if you're just wanting to
>>> analyze it in VMD.
>>> You don't say what it is that you want to analyze, so
>>> it's hard to make
>>> any concrete statements about what you would be missing
>>> if you just
>>> proceeded using the mol2 file. If you can be more
>>> specific about your
>>> goals it would be easier to give further guidance.
>>>
>>> Cheers,
>>> John Stone
>>> vmd_at_ks.uiuc.edu
>>>
>>> On Wed, Oct 01, 2014 at 10:53:14PM +1000, M Karim wrote:
>>>> Hi,
>>>> I have a structure that doesn't have a psf file for
>>>> it. It is not a
>>>> biological material (Nucleic Acid or Protein, so I
>>>> think Automatic psf
>>>> builder does not work for it) I have the xyz, pdb,
>>>> and mol2 as well as the
>>>> trajectory files for it.
>>>> How can we make the psf file for this structure to
>>>> better analyze it in
>>>> VMD?
>>>>
>>>> Thank you
>>>> MK
>>>
>>> --
>>> NIH Center for Macromolecular Modeling and Bioinformatics
>>> Beckman Institute for Advanced Science and Technology
>>> University of Illinois, 405 N. Mathews Ave, Urbana, IL
>>> 61801
>>> http://www.ks.uiuc.edu/~johns/ Phone:
>>> 217-244-3349
>>> http://www.ks.uiuc.edu/Research/vmd/
>>>
>>
>>
>
>
>
>
>
> --
> Best,
> /A
>
>

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:22:55 CST