PDBx/mmCIF is the STAR/CIF-based macromolecular structure format that succeeded legacy PDB as the RCSB Protein Data Bank's primary distribution format. NEP's reader is read-only and covers atomic coordinate, unit-cell, and descriptive metadata categories.

Transparent Access: Read PDBx/mmCIF files with nc_open() after calling NC_MMCIF_initialize().

Category-to-NetCDF Mapping:

_atom_site (looped) → a single atom dimension (file order) and per-atom variables: atom_site_Cartn_x/y/z (NC_DOUBLE), atom_site_id, atom_site_label_atom_id, atom_site_label_comp_id, atom_site_auth_asym_id, atom_site_auth_seq_id, atom_site_occupancy, atom_site_B_iso_or_equiv, atom_site_type_symbol, and atom_site_group_PDB ("ATOM" or "HETATM").
_atom_site.pdbx_PDB_model_num distinct values → a model dimension (1 if the item is absent); coordinate variables are shaped [model][atom].
_cell/_symmetry → global attributes cell_length_a/b/c, cell_angle_alpha/beta/gamma, symmetry_space_group_name_H-M, symmetry_Int_Tables_number. Omitted entirely if the file has no _cell/_symmetry category.
_entry/_struct → global attributes entry_id, struct_title.
? (unknown) and . (not applicable) placeholder values both map to the destination variable's NetCDF fill value; no separate mask/flag variable is added.

Known Limitations: _entity_poly_seq-derived sequence data, _atom_site_anisotrop (ANISOU-equivalent) data, and _struct_conn/_struct_conf/_struct_sheet_range (bonds/secondary structure) are not currently handled. Multi-model (NMR ensemble) mmCIF is untested since all current test files are single-model X-ray structures.

Use Cases: Protein/macromolecular structure analysis, RCSB-sourced X-ray crystal structures.

Enabling:

cmake -B build -DNEP_ENABLE_MMCIF=ON # CMake (default OFF)

Dependencies: None — a custom STAR/CIF tokenizer; no external parsing library is required.

Resources: PDBx/mmCIF Dictionary · RCSB Protein Data Bank

Example:

#include "mmcifdispatch.h"
NC_MMCIF_initialize();   /* register UDF8; safe to call even if already registered */
nc_open("structure.cif", NC_UDF8, &ncid);
nc_inq_varid(ncid, "atom_site_Cartn_x", &varid);
nc_get_vara_double(ncid, varid, start, count, coords);
nc_close(ncid);

Visualization

NEP includes Python visualization examples in examples/viz/ that open PDBx/mmCIF files through the NetCDF UDF interface and write publication-ready PNGs.

plot_mmcif_1j7w.py — 3D scatter of ATOM/HETATM Cartesian coordinates from test/data/mmCIF/1J7W.cif (deoxy haemoglobin beta-Y-Q mutant).
plot_mmcif_2w6v.py — 3D scatter of ATOM/HETATM Cartesian coordinates from test/data/mmCIF/2W6V.cif (deoxy haemoglobin-xenon complex).
plot_mmcif_4hhb.py — 3D scatter of ATOM/HETATM Cartesian coordinates from test/data/mmCIF/4HHB.cif (human deoxyhaemoglobin).

Enable the examples with:

cmake -S . -B build -DNEP_BUILD_EXAMPLES=ON -DNEP_ENABLE_VIZ_EXAMPLES=ON -DNEP_ENABLE_MMCIF=ON

Run only the mmCIF visualizations with ctest --test-dir build -R viz_mmcif --output-on-failure. Generated artifacts are mmcif_1j7w_structure.png, mmcif_2w6v_structure.png, and mmcif_4hhb_structure.png (each with a companion _metadata.txt) in the visualization build directory.