Make ChimeraX fetch 8 character PDB identifiers

[Tom Goddard]

The PDB announced they will start using 8 character PDB identifiers. These identifiers are already in all PDB mmcif files from August 2021 on. Once they run out of 4 character identifiers, any new entries will only be available using 8 character identifiers. And they say those will only be provided in mmCIF format.

​https://www.wwpdb.org/news/news?year=2022#630fee4cebdf34532a949c34

Our current fetch command "open 1a0m" looks specifically for 4 character identifiers and will need to handle 8-character identifier too. Probably should also check that it is alpha-numeric with no "." to avoid mistaking 8 character filenames for identifiers. Also the code may need a change to the fetch url since the PDB also seems to be putting a "pdb_" in front of the 8 character identifier in most places.

[pett]

Well, I have added support for 8-character IDs (​https://github.com/RBVI/ChimeraX/commit/91db3471cdaf07fa1f11788fc9edab06d670e9db) but am mystified as to how to test it, since there are no actual 8-character ID entries. Using the site we currently fetch from (​http://files.rcsb.org/download/%s.cif) works for 4-character IDs but not the equivalent (e.g. "0000" in front) 8-character IDs, even with "pdb_" prepended. Using the DOI brings you to the entry's page -- it does not download a file.

Open to testing ideas. Or should I prod the PDB somehow?

[Tom Goddard]

I guess we need to ask PDB if they have files online with names like pdb_00001xyz.cif that would allow a fetch with the 8 character id. I didn't see them online and I guess they probably don't exist yet. Probably you should ask PDB if those files are available and tell them we want to test ChimeraX fetching the files.

[pett]

Have shot off mail to the PDB asking about how to test 8-character IDs.

[pett]

The RCSB has pledged to support 8-character-ID downloads. I tried the example entries referenced in the link in the ticket description, and ChimeraX fails to read them with "stying lost" a few thousand lines into the file, at the point where the 5-character CCD entry is encountered in the ATOM/HETATM records. I don't know if the actual final form of these entries will behave the same way or if they will allow 5 columns for the residue name throughout. For convenience, I have attached the example files to this ticket.

[pett]

There is already another ticket for lost styling I think. Can't do anything more on 8-character IDs until the RCSB provides a way to fetch them.