[Biopython-dev] [Bug 2619] Bio.PDB.MMCIFParser component MMCIFlex commented out in setup.py

[bugzilla-daemon at portal.open-bio.org]

http://bugzilla.open-bio.org/show_bug.cgi?id=2619





------- Comment #3 from biopython-bugzilla at maubp.freeserve.co.uk  2008-10-22 04:51 EST -------
(In reply to comment #2)
> > In the short term, a note in the documentation would help... were you
> > refering to "The Biopython Structural Bioinformatics FAQ"?
> > http://biopython.org/DIST/docs/cookbook/biopdb_faq.pdf
> 
> The FAQ in part, but there is also a link from RCSB that claims BioPython can
> parse mmCIF:
> 
> http://sw-tools.rcsb.org/

I'll make some documentation updates along the lines of "The Bio.PDB mmCIF
parser isn't installed by default due to cross platform compilation problems",
and see if anyone on the dev mailing list has any bright ideas for detecting
flex at install time.

> I've run the Bio.PDB mmCIF parser over all of PDB, and it plain fails on >10%
> of files (>40,000 files, >5,000 failures, mostly spurious missing key
> exceptions).  From what I've seen, it seems that an inconsistency in one table
> of a mmCIF file throws a wrench in the whole parse.  

Would you mind reporting a separate bug on this (tiny sample script, the
exception error(s), and URLs for a couple of the 5000+ failures)?

> I tried the C++ mmCIF parser from ncbi (only on a few files so far) and it
> doesn't suffer these parse problems (though it reports the faulty entries).

Given the number of PDB problems Bio.PDB has to deal with, its sadly not
surprising that mmCIF files also suffer from this kind of thing.

> If Bio.PDB were to be updated, this seems like a good candidate for a back
> end (assuming its portable).  I have the inclination, maybe not the time ;),
> to do this, unless this should fall to Thomas or others.

I would worry about the cross platform support (in particular Windows), but
also the additional complication to building Biopython.  Its certainly worth
discussing if you or Thomas are keen.

Peter


-- 
Configure bugmail: http://bugzilla.open-bio.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.