Dear Jianhui Tian,
We can post-process data generated by R.E.D. Server/R.E.D. IV or
R.E.D. III.x by using R.E.D. Python. Just send a request from the
R.E.D. Server private assistance service; See
http://q4md-forcefieldtools.org/REDS/faq.php#5.
If you construct ATP from
http://q4md-forcefieldtools.org/REDDB/Project/F-90/ you will see that
the ATP mol2 file (FF lib) is composed of 4 residues; loading a FF lib
composed of >1 residue_s_ in the LEaP program can only be used if the
PDB file used to match the FF lib only contains ATP (i.e. not an
ATP-protein complex for instance). This is a strong limitation in
LEaP. The answer to this problem is to generate a single residue name
for ATP (with obviously the two following rules: two atoms cannot
share the same name in a given residue, and the atom & residue names
in the FF lib have to match these in the PDB file). R.E.D. Python
allows atom typing, complex fragment fusion (to overcome bad
connectivities perform by LEaP at some chiral centers), full residue &
atom name control (a single residue can be requested), a mol3 file
format is also generated; see
http://q4md-forcefieldtools.org/Tutorial/leap-mol3.php, and far more...
regards, Francois
Quoting Jianhui Tian <jianhuitian.gmail.com>:
> Dear Francois,
>
> Thanks a lot for your suggestions. I will try the building block approach.
>
> J.
>
>
> On Tue, Sep 17, 2013 at 3:11 AM, FyD <fyd.q4md-forcefieldtools.org> wrote:
>
>> Dear Jianhui Tian,
>>
>> Looking at this structure took 2 sec to be loaded in vmd, and by
>> simply looking at this ATP analog:
>>
>> The problem you encounter was discussed several times & the answer is
>> known:
>>
>> - problem of convergence? try a more robust program...
>>
>> - can you get an optimized structure? if yes, its conformation will be
>> all, but representative...
>>
>> I would not optimize this type of whole molecule with multiple charge
>> values. Instead I would use the building block approach; i.e. R.E.D.
>> Server & R.E.D. IV and not antechamber in this type of work...
>>
>> See http://q4md-forcefieldtools.org/REDDB/projects/F-90/
>>
>> regards, Francois
>>
>>
>> > Sorry to bother and take your time again.
>> >
>> > The no convergence happens during the sqm calculation for charge.
>> > Is it doable/reasonable to get the topology like this:
>> > 1. calculate the topology for the AMP PNP with -2 charge (this charge can
>> > converge)
>> > 2. get the charges using another way
>> > 3. fix the charges in the final topology file
>> > The assumption I have here is the bonded and vdw parameter will not be
>> > affected by the net charge of the ligand.
>> >
>> > Otherwise, I do not know how to run the antechamber in a
>> > broke-up/step-by-step way.
>>
>>
>>
>> > On Mon, Sep 16, 2013 at 10:13 AM, case <case.biomaps.rutgers.edu> wrote:
>> >
>> >> On Fri, Sep 13, 2013, Jianhui Tian wrote:
>> >> >
>> >> > Please find in the attachment the PDB of AMPPNP I used. Thanks a lot
>> for
>> >> > your time and your kind help.
>> >>
>> >> This is a problem that requires more time than I have had so far. It
>> may
>> >> be a
>> >> problem with the initial geometry. What I've found so far is that
>> >> Gaussian09
>> >> also fails with this input, and that PM6 seems OK, and can optimize the
>> >> geometry fine. I've not yet had time to investigate other options; you
>> >> might
>> >> try geometry optimization at the PM6 level, then try using that for AM1
>> >> calculations.
>> >>
>> >> Of course, in general, using gas-phase calculations for a -4 anion is a
>> >> stretch. You might consider adding a proton and looking at the -3
>> state,
>> >> since the pKa of ATP is around neutral pH.
>> >>
>> >> ...dac
>>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Sep 17 2013 - 23:00:03 PDT
