[o-info] MOI v11 feature in pdb_read & some loose talk

alwyn@xray.bmc.uu.se

i've got another v11 'feature' to report, courtesy of lothar esser.

if in pdb_read you want cns segid's, do not want to concatenate the  
residue name and segid, and if the chain and seg IDs are the same 1  
character, THEN on pdb_write you will loose the chain id character.

i've already fixed it and will recompile things over the weekend.

i have had a look at what the ePDB is planning this summer, http:// 
remediation.wwpdb.org/
i checked the various test PDB files there too. overall, i think the  
biggest hiccup will come from the atom name changes and nucleic acid  
residue type changes. i think this is GOOD, but it will need updating  
of my dictionaries. i don't know if cns, ccp4 are planning a switch  
to fit in with the wPDB timetable. it could be a period of  
'interesting' times with conflicting coordinates and dictionaries.

the remediated (nice word) files have a remark record in them  
indicating that they have the new changes. i intend to make use of  
this, and PDB files that O generates will contain a similar record.

as you probably know, o forms a 6 character residue name out of the  
contents of columns 22-27, removing spaces, capitalizing and shifting  
left. this has been called 'a gross simplification to the pdb format'  
but i still sleep moderately well at night. in v11 i tried to help  
people working on structures with very many chains by allowing the  
concatenation of residue names with SEGIDs. this, of course, causes  
problems when trying to cram long SEGIDs and names into just 6  
characters. wPDB are not planning on using SEGIDs, but they've put  
out the suggested use of PART/PRTEND cards for discussion. i could  
support this by generating separate O-style molecules for everything  
between each set of PART/PRTEND cards. anyone got opinions on this?
how about creating a new molecule each time the CHAINID not equal to  
SEGID? i could support this with some limitations on SEGID width to  
help form the molecule's name, or by some other naming ploy.

Alwyn Jones
alwyn@xray.bmc.uu.se
http://xray.bmc.uu.se/alwyn

_______________________________________________
o-info mailing list (moderated)
o-info@o-info.bioxray.dk
http://o-info.bioxray.dk/mailman/listinfo/o-info