Re: CCL:I/O in Charmm
I'd like to add a couple additional comments, from the perspective of an
experienced user of the academic version of CHARMM (as opposed to MSI's
CHARMm). My recommendation to the original poster was to create the
phosphorylated ribose topology (RESIdue) by hand editing a copy of a
nucleic acid topology file, paying careful attention to atom parameter
types. This is probably the most reliable way, assuming all the needed
parameters are available (they should be, for phospho-ribose).
On Sun, 3 Sep 2000, Don Gregory wrote:
> Really, there are two separate issues here; (a) using InsightII to create
> CHARMm input files, and (b) what to do when a chosen force-field
> is missing parameters.
Since the original question referred to one of the academic topology
files, a potential third issue is using mixed parameters; in general,
it's a bad idea. RTF/PARAM files are developed and must be used with
explicit pairings, and there are several sets of these files available
today from various development efforts, both academic and commercial.
Further, one should not "borrow" parameters and RTF files from one
development effort for use with another, at least not without detailed
testing. One fundamental difference for the academic parameters is that
they assume less transferrability than e.g. the commercial sets
available from MSI with Quanta and InSightII.
> The easy one first: Indeed, InsightII can write CHARMm PSF and
> RTF files. Whenever your research leads you to modify
> the (more or less) 'standard' molecule/residue types that
> are provided in the RTF files that come with InsightII, we
> heartily recommend you use the PSF creation capabilities
> of InsightII to make the 'whole-system' unique molecular
> topology description that CHARMm needs, i.e. the PSF file.
> That is what this capability is intended for.
As noted in Don Gregory's reply, this assumes the use of the CFF
parameters, one of the several development efforts. As I note above,
one shouldn't try to combine a CFF-derived molecular description (RTF,
PSF) with e.g. the academic parameters w/o testing. I agree
whole-heartedly with Don Gregory that any such extension to an existing
or default topology and parameter set should be documented in any
publications using the changes, along with how the changes were tested.
Some journals even insist that you do so.
In general, caution and attention to detail are needed when adding new
molecules to CHARMM, especially when using the academic parameters.
Even more work is needed if it turns out that there are missing
parameters. Although it may be possible to point-and-click your way to
a molecular description and parameters w/o any gross flaws, I recommend
a detailed comparison to existing similar residues before going too far
with any detailed molecular mechanics studies.
A final note-- my personal experiences with programs such as ChemNote
(supplied w. Quanta) for building new CHARMM residues has generally been
unsatisfactory. I haven't tried the InSightII tool (we don't use CFF).
Original question:
> At 06:37 PM 9/3/00 -0400, Xiang(Simon) Wang wrote:
> >Can someone help me on how to read small molecule structure in Charmm?
> >For example, phosphoribosyldiphosphate (PRPP), which can not be found
in
> >top_all22_model.inp file. I could get the PSF file from InsightII. But
> >when running, Charmm gave the warning of lacking of parameters. Can
> >someone help me out of it?
--
Rick Venable =====\ |=| "Eschew Obfuscation"
FDA/CBER Biophysics Lab |____/ |=|
Bethesda, MD U.S.A. | \ / |=| ( Not an official statement or
Rick_Venable ^at^ nih.gov | \ / |=| position of the FDA; for
that,
rvenable ^at^ speakeasy.org \/ |=| see http://www.fda.gov )