Nathaniel Echols wrote:

On Wed, Apr 20, 2011 at 9:33 AM, Zhang yu wrote:

...

When I run MR in Phenix, I found that there is a numbering error in the solution PDB file

The error is that Phenix numbers residues from 0 again when the residue numbers are over 999, though they belong to the same chain.

For example

ATOM 11513 NE2 HIS C 999 -35.122 54.953 55.285 1.00 69.80 N ATOM 11514 N MET C 0 -36.544 51.378 60.309 1.00 64.17 N ATOM 11522 N VAL C 1 -37.840 53.015 63.272 1.00 69.46 N

Does anyone has idea about it?

Phaser has its own PDB parser, and it looks like it only reads characters 2-4 in the residue number. This sounds like a bug, but there may be some reason for it - we'll check this out.

There was discussion a few years back about an extended pdb format in which the fourth digit would go not just 0-9 but 0123456789ZBCDE...XYZabcdefghijkl....z (and the other digits would also if the fourth digit was greater than 9) So maybe the code that handles that isn't kicking in where it should? The folks who want to deposit structures of whole bacteria will have to wait a little longer. ed

I just realised that I hadn't reported the outcome to the BB. There was a bug in the format used to write PDB files in Phaser. Airlie McCoy fixed it on Thursday, so nightly builds since then will not have this error. Best wishes, Randy Read On Apr 20 2011, Nathaniel Echols wrote:

On Wed, Apr 20, 2011 at 11:45 AM, Edward A. Berry wrote:

...

There was discussion a few years back about an extended pdb format in which the fourth digit would go not just 0-9 but 0123456789ZBCDE...XYZabcdefghijkl....z (and the other digits would also if the fourth digit was greater than 9) So maybe the code that handles that isn't kicking in where it should?

This is Ralf's "hybrid36" format. I think Phaser will use that encoding for atom records, but not residues. In any case, using the hybrid36 format is unnecessary in this case, since the residue number field is officially (according to the PDB) four characters long.

...

The folks who want to deposit structures of whole bacteria will have to wait a little longer.

The PDB won't even handle ribosome structures without splitting them into multiple files, so this has been a problem for at least a decade. (Phenix does not have this limitation, of course, nor do Coot, Refmac, or CNS, as far as I know.) Residue numbers are less likely to overrun the official format, but I think there are structures for fragments of proteins where the actual residue numbering starts above 9999. I'm not sure what the PDB does to these.

-Nat _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb