Hi Paul, Many thanks for getting back to me on this.
I can confirm that the 6ttq mmcif also fails to open on my version of coot, while the PDB opens fine. On the links issue – another deposited example is 6r4t; while these are admittedly metal coordination links rather than covalent links, I believe the PDB links / mmCIF struct_conn loops are of the same formats. For the PDB file, all metal coordination distances are shown by coot as dotted lines, whereas coot fails to draw these for the mmCIF file. If you’re about to release a new coot version that solves either or both of these issues (in particular, does it move from mmdb2 to GEMMI?), then I’ll happily proceed with deposition. Will this new version be part of a future CCP4 release? Thank you for all your help! Best, Neil From: Paul Emsley <pems...@mrc-lmb.cam.ac.uk> Date: Saturday, 14 October 2023 at 17:51 To: Neil Rzechorzek <neil.rzechor...@crick.ac.uk>, "COOT@JISCMAIL.AC.UK" <COOT@JISCMAIL.AC.UK> Subject: Re: Coot/mmdb2 issues with NCS and links in mmCIFs External Sender: Use caution. On 13/10/2023 13:12, Neil Rzechorzek wrote: Dear colleagues, Hi Neil, I have been iteratively rebuilding and refining a crystal structure using coot (v. 0.9.8.91) and Phenix. Refinement makes use of NCS restraints, and the structure additionally has some covalent links between protein sidechains and bound ligands. OK, sounds interesting. NCS parameters and links are output by Phenix in both PDB and mmCIF file formats. Coot happily opens the PDB file, and the covalent links are displayed as the usual dotted lines. However, coot fails to open the mmCIF file. According to an old ccp4bb post (https://www.jiscmail.ac.uk/cgi-bin/wa-jisc.exe?A2=ind1910&L=CCP4BB&D=0&P=73942), this appears to be an mmdb2 library issue that I gather has not been resolved. That's right. But there's more. I've been working on a new version of Coot and that reads 6ttq (for example) just fine. In particular, the _struct_ncs_oper category in the mmCIF is the culprit and - as the previous poster noted - removing this whole section from the file enables coot to open it. I gather this is akin to deleting the MTRIX section in a PDB file, though am not sure whether this is either permitted or advisable for a deposited structure. Indeed. Additionally, coot doesn’t show the sidechain-ligand links in this modified mmCIF file. The correct _struct_conn.id loops are present – is there a switch somewhere that encourages coot to show them? There is no user-control over this. I’ve checked other deposited mmCIF files with such links, and they too suffer from the same problem. Do you have a particular code in mind? As I’m now depositing the structure at the PDB, I’m trying to resolve these issues prior to accepting the final annotated version. While end users will always be able to open the PDB file in coot, mmCIF seems to be the preferred file type for many. This will increasingly become the case. Apologies if I’ve missed other posts/solutions to these issues, or have misunderstood something – I did search, but could not find answers. Any suggestions would be extremely welcome! The ball is in my court. The version of Coot that I am using (from the main branch of the Coot repo) doesn't have these problems, AFAICS. I need to get it in shape for release. Since mmdb2 is managed by the CCP4 team, I will likely post this to ccp4bb also. They will say "if mmdb2 is the problem then GEMMI is the solution" - or some such. Paul. The Francis Crick Institute Limited is a registered charity in England and Wales no. 1140062 and a company registered in England and Wales no. 06885462, with its registered office at 1 Midland Road London NW1 1AT ######################################################################## To unsubscribe from the COOT list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=COOT&A=1 This message was issued to members of www.jiscmail.ac.uk/COOT, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
