On 25/11/10 15:25, Phil Evans wrote:
This is a coot/refmac problem, but mainly coot I think, and I didn't want to
cross-post.
I'm deeply confused about atom& residue naming for nucleic acids.
There are several groups here working with RNA, and I recently broke coot for them, by installing the latest
"experimental" refmac, and more particularly it's new dictionaries, into directory $CLIBD_MON. These
dictionaries use what I believe are the most recent naming conventions, ie RNA residue type are A, G, U, C (not Ar
etc), and DNA residues are DA, DG, DT, DC. Also ribose atoms are named with "'" rather than with
""*".
OK, I didn't know that.
Thus AFAICS there are two name sets here
1. "Old" names: residue names Ar, Gr, Ur, Cr for RNA and * in the ribose atom
names.
2. "New" names: residue names A, G, U, C for RNA and ' in the ribose atom names.
The coot script (in COOT/bin/coot) by default picks up dictionaries from
$CLIBD_MON (with new names now), unless I uncomment the line
COOT_REFMAC_LIB_DIR=$COOT_PREFIX/share/coot/lib
(and the corresponding EXPORT command)
in which case it uses Coot dictionaries (in COOT/share/coot/lib/data/monomers),
which use old names
agreed.
There are then the following cases:
1) if I leave Coot pointed at the (new) refmac dictionaries, then
regularisation fails, both with files with old names (since they don't match
the dictionaries) and with a file which I've edited to the new name convention
(since Coot automatically and inconveniently converts a new-name file to old
names)
2) if I point Coot to its own dictionaries, then files with both old& new name
conventions work, but new names are converted to old names, so output files will be
like that
Refmac with the new dictionaries seems to be able to read RNA files with either
new or old naming, and preserve the naming scheme.
So my question is: how do I set up a system which will work with Coot, Refmac,
and preferably Molprobity and Phenix too? What are we _supposed_ to do?
I hope that this will do the trick (in your ~/.coot (or some such)):
(set-convert-to-v2-atom-names 0)
(I see that this is not documented, sorry about that).
1) Are the atom naming conventions described anywhere in an intelligible
fashion? I found it hard to find things on the RSCB web site, particularly for
hydrogen atom names
:-/ we are in the same boat.
2) the "new" dictionaries from the Refmac site seem to have the old (v2?) hydrogen names
for amino-acids, at least as far as I can tell from the hydrogen names given by Molprobity which I
believe are in the "new" v3 convention (see PS point 1)
That was my understanding too.
Paul.
