OBSLTE from pdb and cif input
rimmartin opened this issue · 6 comments
Hi @wojdyr and everyone,
Should OBSLTE be added to Structure::info as the cif tag _pdbx_database_PDB_obs_spr or do you think the record(can be multi-line) should be parsed in more detail at the read step?
If Structure::info would the pdb and cif wind up being the same info to interpret by app?
Thank you
Note that your example has SPRSDE instead of OBSLTE.
I looked at _pdbx_database_PDB_obs_spr in the current PDB database, and unfortunately in a few cases
(namely: 3EZB 5HUZ 3J7O 3J7P 3J7Q 3J7R 1NEW 9PAP 2R2X 1SAF)
it contains a few entries, for example
loop_
_pdbx_database_PDB_obs_spr.id
_pdbx_database_PDB_obs_spr.date
_pdbx_database_PDB_obs_spr.pdb_id
_pdbx_database_PDB_obs_spr.replace_pdb_id
_pdbx_database_PDB_obs_spr.details
SPRSDE 1987-01-15 9PAP 8PAP ?
SPRSDE 1986-10-24 9PAP '3PAD 8PAP' ?
although in the corresponding PDB it's only one line
SPRSDE 24-OCT-86 9PAP 3PAD 8PAP
I didn't check how it looks in the obsolete PDB entries.
Out of curiosity, what would you use it for?
Ah, the https://www.rcsb.org human interface by default jumps 1qon to its replacement 6xyu.
I'm being asked to provide the info and choice to users before they do a large calculation when we pull from the web service programmatically(it doesn't auto-switch to the more recent). For example a user sets our software to modeling missing loops and refining a number of loops to see if they now fit the density. When there could be a better model in the rcsb that already has the loops. Or better density or resolution.
pdb 1qon
OBSLTE 15-FEB-23 1QON 6XYU
mmcif 1qon
_pdbx_database_PDB_obs_spr.id OBSLTE
_pdbx_database_PDB_obs_spr.date 2023-02-15
_pdbx_database_PDB_obs_spr.pdb_id 6XYU
_pdbx_database_PDB_obs_spr.replace_pdb_id 1QON
_pdbx_database_PDB_obs_spr.details ?
#
Perhaps the web service that you use can return status of the entry without downloading a file?
Or maybe it'd be more convenient to use the list from:
https://files.wwpdb.org/pub/pdb/data/status/obsolete.dat
Thanks for pointing out obsolete.dat
I'm using http://files.rcsb.org/download/ which I think mirrors the wwpdb
They want both ways; it happens when you give them options:-)
If a user already has pdb's what we're trying to do is be nice to their budget and not run unless they know about an updated structure
OK, I'm not against it.
Are there any examples of multi-line OBSLTE?
Or examples when this record has multiple new PDB IDs (in either PDB or mmCIF format)?
I'll keep an eye out. Will ask those who run large sets for an example