VMD-L Mailing List

From: JC Gumbart (gumbart_at_ks.uiuc.edu)
Date: Wed Feb 29 2012 - 20:36:18 CST

You're going to need to be a lot more specific about what you did and how it failed if you want help. The first thing you'll need, for example, is a topology file. I think Molefacture can help, but you should also look at other examples in the force field.

On Feb 29, 2012, at 5:52 AM, Francesco Pietra wrote:

> I would like to try FFTK to parameterize a simple complex of a
> transition metal. How to setup the "Input PSF File"? Perhaps my
> attempt through the autoPSF plugin is not the right way as, obviously,
> it fails to provide the desired file.
>
> thanks
>
> francesco pietra
>
> On Sat, Feb 18, 2012 at 11:39 PM, JC Gumbart <gumbart_at_ks.uiuc.edu> wrote:
>> I will add that this is similar to what I did for an Fe ion coordinated by a
>> protein. Obviously you won’t learn anything about binding/unbinding, but it
>> should at least represent the bound state reasonably well, vs. the
>> alternative of having the ion free (which may be interesting to look at for
>> comparison anyway). One additional note though: I had to create some new
>> atom types for coordinating residues in order to have unique parameters for
>> all the bonds/angles/dihedrals. Then I had to copy existing protein
>> parameters for those types to the new ones.
>>
>>
>>
>>
>>
>> From: owner-vmd-l_at_ks.uiuc.edu [mailto:owner-vmd-l_at_ks.uiuc.edu] On Behalf Of
>> Mayne, Christopher G
>> Sent: Saturday, February 18, 2012 3:40 PM
>> To: chiendarret_at_gmail.com
>> Cc: vmd-l_at_ks.uiuc.edu
>> Subject: vmd-l: Re: About FFTP with transition metal complexes
>>
>>
>>
>> Francesco,
>>
>>
>>
>> FFTK was designed primarily with small molecule organics in mind. I am not
>> familiar with the specifics of parameterizing metal complexes for CHARMM;
>> however, if it follows the same (or similar) workflow as organics, then FFTK
>> should, in principle, be applicable/useful.
>>
>>
>>
>> To answer your questions directly:
>>
>>
>>
>> 1) FFTK relies heavily on QMtool for parsing Gaussian log files within the
>> charge optimization routine. Try loading the DFT calculation log file into
>> VMD using QMtool; if the optimization steps are properly loaded and you can
>> see step energies pushed to the TkConsole during the load, then it *should*
>> work in FFTK.
>>
>>
>>
>> Note 1: you can still use FFTK to setup the calculation by changing the
>> Route details with in the "Gaussian Setting" section using the appropriate
>> Gaussian keywords
>>
>> Note 2: make sure that you calculate separate single point energies for the
>> metal complex and a TIP3P water molecule at the same level of theory as used
>> in the water interaction calculation.
>>
>>
>>
>> 2) This is a more of a general psf/topology question and not directly an
>> FFTK question. Since I don't work with metalloproteins, I'm only guessing
>> here; hopefully someone else will chime in.
>>
>> You should be able to leave all topologies in tact and write a patch to
>> adjust the partial charges for the ligating residues in addition to adding
>> bonds/angles/dihedrals necessary to describe the ligated complex.
>>
>>
>>
>>
>>
>> Chris
>>
>>
>>
>>
>>
>> Date: Sat, 18 Feb 2012 07:23:01 +0100
>> From: Francesco Pietra <chiendarret_at_gmail.com>
>> Subject: vmd-l: About FFTP with transition metal complexes
>>
>> My aim is to parameterize a transition metal complex for a
>> metalloproteins. Two questions about FFTP, a most waited tool which
>> going to foster a lot of research work.
>>
>> (1) Calculation of partial charges. When a transition metal is
>> involved, a single-point HF would be inappropriate. Current literature
>> about any type of dealing with such complexes suggests to go on - from
>> beginning to end - with DFT, with higher basis set for the metal or
>> metal ion. Who has specific experience for CHARMM ff to this regard.
>> In any case, is FFTP prepared to accept the results (partial charges,
>> geom data, force constants) from a DFT type of calculation? If so I
>> could try to apply the plugin to my case.
>>
>> (2) As there is no tutorial yet for FFTP, it is not yet clear to me
>> how to bind the parameterized metal complex to its protein. I make an
>> example: for a metal ion bound to HIS and ASP residues, I make a model
>> including a portion only of the HIS and ASP residues, while carrying
>> out partial charge and bond data (geometry and force constants) for
>> that model. Correspondingly, I have a protein from which the portion
>> included into the metal complex has been removed. What about the
>> procedure to establish chemical bonds between the abridged protein and
>> the metal complex? (i.e., the equivalent of the command "bond xxx
>> yyy" with Amber's LEaP).
>>
>> Thanks
>> francesco pietra