Part of the CCP4 Developers meeting
Update the ccp4 monomer library use acerg derived dictionaries. Using ligand dictionaries taken from the CCD.
Now we are assesssing the quality of models by looking at the results of refinement.
Imminent first update with the "easy wins" - verified that give good refinement. Widely criticised "too conservative" selection criteria.
Sameer mentioned CSD entries distributed by the PDBe. How are we doing when testing against these? No clear answer.
AP: Rob to address this at the next ligands meeting.
Noted technical issues of cif formatting - including link information
Monomer library is ready for release, revision 107.
Sugars will be addressed separately.
Discussion: how to do a search of the monomer library for a new molecule (say a SMILES string). Maybe by graph seraching - maybe by canonical SMILES? Using fingerprinting for faster searching. What is the use case? Unresolved.
Discussion: When will the next round be done - syncing with ccp4 releasse?
"Monthly updates" says Rob. That seemed acceptable. Let's look at the progress next ligand meeting.
AP: RN to progress with the release of monomer library updates.
Internally, it makes "combo" ligand
execute using:
acedrg -L instruction-file.txt
Discussion: of details of transfer of link information in the cif dictionary.
Discussion: struct conn in the mmcif files
AP: needs to be output by cif by Coot - PE
Discussion: I want to be able to name the link (id) Keyword in link-instructions.txt. This will be a topic this at next ligand meeting.
AP: PE and Fei
Sameer to talk about progress of link information in PDB deposition and publication.
Atom name matching now available in the Make Ligand now
Better output for matching when dictionary not found.
Briefly mentioned Coot's new N-linked carbohydrate builder
Note: we are using latest CCP4 expert version.
XChem example. Jiggle fit used for fitting. Everything worked straightforwardly.
AP: PE to add "Jiggle Fit" to the Ligand menu
Ugly looking plots for data (ice ring). Auspex is useful in this case - identifies ice rings. Martin: Auspex output should be available as part of the import task report page.
Discussion: We want a menu item so that you don't have to type the 3-letter code when importing monomer into Coot after it's been created in CCP4i2.
AP: Paul/Martin to add the button.
Stuart: Make a "extensions looker-upper" as a dropdown option in the Coot interface, so that people don't have to find/download/run scripts.
Paul: That's a good idea.
Dave: Not clear that users should use the link maker after importing a ligand. If there's a clash then they may think they have to manually resolve the clash (remove an atom) before making the link. The correct path is not intuitive.
Problem encountered with having to delete hydrogens manually.
Judit: If the user deletes the wrong hydrogen it'll cause errors downstream.
Dave: Involved process. If it goes wrong, when does the user find out that it goes wrong?
AP: Paul to see if Coot can apply the atom deletions in the dictionary.
Discussion: Issue with links between different chains. wwPDB allows it (but changes chain-ids so that it's not necesary). Refmac allows it. Coot doesn't.
AP: Coot needs to allow it - PE to investigate.
The ligand from the PDBe is not the ligand that the authors soaked into the crystal - missing vinyl double bond: 4zzm. The missing bond makes the generation of the link easier than it should have been.
Martin: how about doing screen capture movies of creating links?
Dave agreed: Process is logical, once you've seen it. But new users will be confused.
AP: PE to investigate making some videos.
Dave: I see a limitation: if I want to modify a ligand after creating a link, I've got to go through the whole process again.
Paul: Yes.