[gmx-users] Re: lipid.itp and grompp

[Marc Baaden]

>>> Thad Harroun said:
 >> A. Including a second [defaults] superceeds the first [defaults] and gives
 >> a grompp warning.

Yes. SHould rather be a fatal error, as if you were changing combination
rules you could get funny results.

 >> B. A second [atomtypes] section by itself (without subsequent
 >> [nonbond-params]) is okay, and any new non-bonded parameters are
 >> calculated based on the second [defaults] section. The previously read
 >> [nonbond-params] section is retained.(?)

No ! A second [atomtypes] is not ok ! Check it by changing any of the
previous nonbond_params, which will have no effect !

 >> C. A second [nonbond-params] section (typically to include the new atom
 >> types) will cause the first to be ignored, and any non-bonded parameters
 >> not in the explicitly stated second [nonbond-params] section are
 >> calculated based on the second [defaults] section. No grompp warning.

I think that's right.

 >> D. B and C also applies to [pairtypes] section.

I don't know :(

 >> The best solution is to merge all [*types] sections under one [defualts]
 >> section.

Yes, and to only have ONE single occurrence of atomtypes and nonbond_params in
precisely that order in your whole topology.

Best wishes,
Marc
-- 
 Dr. Marc Baaden - Laboratory of Molecular Biophysics, Oxford University
 mailto:baaden at smplinux.de  - ICQ# 11466242 -  http://www.marc-baaden.de
 FAX/Voice +49 697912 39550  -  Tel: +44 1865 275380  or  +33 609 843217