From: Kenno Vanommeslaeghe (
Date: Wed Sep 18 2013 - 16:52:41 CDT

Hi Tee,

On 09/18/2013 04:45 PM, Teerapong Pirojsirikul wrote:
> Personally, I am working on an RNA system. Anyway, I have 2 questions
> here. 1) I don't think (but need your confirmation) that my modified
> toppar_water_ions.str should cause any problem for systems with
> regular nucleic acids, right? Those NBFIX terms I have commented out
> have some thing to do with the interaction between Na ion and
> carboxylate oxygens.

That is correct; for a system containing *only* nucleic acids (with
solvent and ions of course), your file is perfectly OK. Just be careful if
you or any of your group members decide to simulate something else ;)

> 2) If I leave those NBFIX as is, I will get an error regarding unable
> to find vdw parameter of OC atom type. My system itself has nothing to
> do with this atom type and when I tried commenting those NBFIX lines
> out, the problem solved. Is this because I haven't loaded what are
> called 'the main files' before using the toppar_water_ions.str file?
> And by that definition, what do those main files exactly mean?

The link in my previous message is to be taken in the context of the
CHARMM General Force Field for drug-like molecules (CGenFF). In that
context, the main parameter file is par_all36_cgenff.prm , as opposed to
the specific stream file for the drug-like molecule (or cofactor) of
interest, molecule.str . The stream file depends on the main file being
read first. A similar situation exists for the other parts of the force
field; stream/toppar_all36_prot_heme.str depends on the main protein
parameter file par_all36_prot.prm being read first.

As for toppar_water_ions.str, it was originally meant to be read after the
main files of the parts of the force fields the user is planning to use,
and that works in CHARMM. But in NAMD, the parameter reader crashes when
encountering NBFIX lines with atom types that haven't been defined before,
so the only way to read an unmodified toppar_water_ions_namd.str is by
reading all of the protein, lipid, carbohydrate and CGenFF main files
first, even if you don't need them. Reading all the files like that was
not trivial in the previous release of the CHARMM force field, but has
become very easy with the new modular version.

In your isolated case, you can just keep using your modified file; reading
all the files has no advantage for you as far as the energy function is
concerned. However, more generally spoken, I would strongly encourage
reading all the files; having a modified file that is not generally
applicable lying around in a file system is a huge liability. Murphy's law
dictates that someone eventually will pick it up and use it in a context
where it's not appropriate - in fact, we almost saw that happening right
here, if Ana would have used your file...



This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:41 CST