Hi all,
so far, we had a lot of theoretic discussions going on concerning the selection topic. Maybe things clear a little if I contribute some concrete examples from my 'PDB is hell' list. Under the bottom line I found out that if you consider HETATM and ATOM records to be different entities, as does PDB, the residue numbers become unique (except for a few real erroneous cases).
* 1ECG : Chain 'A' starts with 'CYS1', but after the TER record (don't trust these records!) there is a HETATM record 'ONL1' which is a modified peptide.
* 2F58 : Chain 'P' has increasing residue numbers till 324 but then continues with residue number 10 to finally end with 12.
* 1CJ0 : (A special friend of mine). Chain 'A' starts with 'TRP-8' increases the numbering till 'GLN-3', followed by 'MET1', behaves normally till 'GLY91' which is followed by 'VAL911' (this is not an insertion code!) to be followed by 'ASN92'. There are a couple more of these strange numbers in this file.
* 1QAW : There are a lot of 'TRP81' residues in the HETATM section, I guess this is due to TRP being a ligand in this case.
* 1BE3 : Chain 'A' has residue number 3 twice: Once where it should be, at residue 'THR3', but then it is found between 'GLY202' and 'GLU204' as 'LEU3'. I assume this is a spelling error and should say 'LEU203'.
These are the most extreme examples of pathologic PDB files that are floating around. I have not checked if this misnumbering is also contained in the CIF files of PDB. (PDB told me that the archive .ent.Z files will not be modified, only their internal CIF equivalents).
I would expect (and RasMol behaves like that) that in case of 1TON, a selection like 'select 53-58 and amino' selects only the NULL chain's peptide residues and a 'select 53-58 and hetero' or a 'select 53-58 and HOH' selects the HETATM records. Care has to be taken for HETATM records in the peptide chain, where a 'select 15-17 and amino' will exclude the 'SME16' residue in PDB file 1GKF.
Hope this helps a little. Chris ------------------------------------------------------- SF.Net email is Sponsored by the Better Software Conference & EXPO September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf _______________________________________________ Jmol-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jmol-users
