IHM Meeting Summary Oct 25 2017

IHM Meeting Summary Oct 25, 2017

Attended by Ben Webb, Tom Goddard, Lukas Hoenig, and Brinda Vallat

1. Atomistic models of HSA are now available on GitHub.

   • ihm_external_files.file_format can be added to address the *.dms file extension for predicted contacts file on Zenodo.
   • ihm_cross_link_restraint.restraint_type can be added in order to visualize restraint violation on ChimeraX.
   • Chimera can visualize the atomistic models, however ChimeraX visualization of atomistic models is in progress.

2. The mmCIF Dictionary Suite does not compile on Ubuntu 17. Lukas to provide the error log file.

3. The enumeration list in ihm_dataset_list.data_type should be made consistent. At present, some data types are precise (e.g., 3DEM volume, 2DEM class average, EM raw micrographs), whereas others are not (e.g., NMR data, SAS data, EPR data). The main reason for this inconsistency is the lack of examples regarding restraints derived from different kinds of experiments. We therefore do not have the appropriate restraint tables defined in the dictionary that can provide precise information regarding the different data types used in I/H modeling. However, some internal consistency regarding the granularity of the data type is preferred. For NMR and SAS, this may be inferred from different restraint types defined in the NEF and the SASCIF dictionaries. For other experiments, we may need more examples and/or input from experts in the field.

   The implementation can involve splitting the ihm_dataset_list.data_type term into experiment type (NMR, EM, SAS etc) and restraint type (NMR-derived distance restraints, 3DEM volumes, 2DEM class averages etc) or retaining a single term and modifying the enumeration list to be internally consistent. The latter may be more appropriate since the details of the experiment itself will be contained in the corresponding experimental data repository.

4. Expanding the atom_site table in the I/H methods dictionary to include atom_site.ihm_model_id causes problems while using visualization software that are capable of reading the atom_site table but cannot handle the additional data item crucial for I/H models. It also causes problems while reusing existing mmCIF parsers that read the atom_site table. One solution is to leave the atom_site table as is and make use of the parent-child relationship between ihm_model_list.model_id and atom_site.pdbx_PDB_model_num, or if needed, create a new table that maps atom_site.pdbx_PDB_model_num to ihm_model_list.model_id.

5. Next meeting is scheduled for Thursday, November 2nd, at 12:00 PM Eastern time.