Legacy PDB (Protein Data Bank) is the fixed-column text format historically used by the RCSB Protein Data Bank for macromolecular structures (proteins, nucleic acids). NEP's reader is read-only and covers coordinate, model, and unit-cell metadata records.

Transparent Access: Read legacy PDB files with nc_open() after calling NC_PDB_initialize().

Record-to-NetCDF Mapping:

ATOM/HETATM records → a single atom dimension (file order) and per-atom variables: atom_site_Cartn_x/y/z, 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").
MODEL/ENDMDL blocks → a model dimension (1 if no MODEL records are present); coordinate variables are shaped [model][atom].
CRYST1 → global attributes cell_length_a/b/c, cell_angle_alpha/beta/gamma, space_group_name_H-M, symmetry_Z. Omitted entirely if the file has no CRYST1 record.
HEADER/TITLE/COMPND/SOURCE → global attributes idCode, classification, depDate, title, compnd, source.

Known Limitations: SEQRES sequence data, hybrid-36 encoded serial/residue numbers, and multi-character chain IDs are not currently handled.

Use Cases: Protein/macromolecular structure analysis, RCSB-sourced X-ray crystal structures, NMR ensembles with multiple models.

Enabling:

cmake -B build -DNEP_ENABLE_PDB=ON # CMake (default ON)

Dependencies: None — a custom fixed-column line reader; no external parsing library is required.

Resources: PDB Format Guide · RCSB Protein Data Bank

Example:

#include "pdbdispatch.h"
NC_PDB_initialize();   /* register UDF7; safe to call even if already registered */
nc_open("structure.pdb", NC_UDF7, &ncid);
nc_inq_varid(ncid, "atom_site_Cartn_x", &varid);
nc_get_vara_float(ncid, varid, start, count, coords);
nc_close(ncid);