Smarter handling of ATOM/HETATM column 21

[p.qian@…]

Column 21 of ATOM/HETATM records is not officially assigned a meaning and currently is interpreted as the fourth character of the residue name (useful for Amber-compatible terminal residue names). COOT (and possibly other programs) use that column as part of a two-character chain ID instead. The proposal to try to allow for that is:

On input: if that column produces consistent 2-character chain IDs with each TER-card-delimited section, then treat as part of chain ID, otherwise as part of residue name (moot if column is all spaces of course).

On output: If multi-character chain IDs in use, use as chain ID, otherwise as part of residue name.

[Tristan Croll]

This is an innovation that was introduced by PHENIX some time back, part of their "Hybrid-36" format - see ​http://cci.lbl.gov/hybrid_36/.

[Eric Pettersen]

Refine the "on output" part to: if possible map the 2-character chain IDs to unique 1-character chain IDs (in case a subset of a larger structure is being written), otherwise use the 2-character chain IDs

[p.qian@…]

Hi,

Does it mean that current version of chimerax accept two-letter ID in Pdb
format? Perhaps you already noticed that both coot and Phenix accept this
format now.

Best regards

Pu Qian

On Wed, 17 Jul 2019 at 21:02, ChimeraX <ChimeraX-bugs-admin@cgl.ucsf.edu>
wrote:

[Eric Pettersen]

Nope, sorry, still on the "to do" list for now...

[p.qian@…]

Hi Eric,

Thanks for your email and for responding to my request on the 21 column for
two-letter ID. I noticed that ChimeraX accepts two-letter chain IDs. How
about Chimera?

Best regards

Pu Qian

On Mon, 15 Jun 2020 at 18:15, ChimeraX <ChimeraX-bugs-admin@cgl.ucsf.edu>
wrote:

[Eric Pettersen]

Hi Pu Qian,

Chimera only allows multi-character chain IDs if the input is mmCIF, not PDB -- which is what this ticket is about. Chimera is probably never going to support multi-character chain IDs because the assumption that chain IDs are a single character occurs throughout Chimera's code. That's one of the reasons we wrote ChimeraX -- to handle modern-size structures better.

--Eric

[Eric Pettersen]

Tomorrow's build will have code to read/write two-character chain IDs in PDB files.