PDBformat

PDBformat is a plain-text file format used for representing three-dimensional structures of biological macromolecules, primarily for data exchanged with the Protein Data Bank. It originated at Brookhaven National Laboratory in the 1970s and remains widely used because of its readability and broad software support, despite newer formats available. A PDB file encodes atomic coordinates, residue identities, and connectivity, along with metadata such as titles, authors, and experimental details.

A PDB file consists of records, each occupying a line with fixed-width fields. The most common records are ATOM and HETATM, which carry Cartesian coordinates of each atom (x, y, z in angstroms), along with occupancy and temperature factor. Each ATOM/HETATM line includes the atom name, residue name, chain identifier, residue sequence number, and an optional insertion code. Residue names use standard three-letter codes, and the file may also indicate alternate locations for atoms. In addition to atoms, PDB files may include TER records to mark chain termini, MODEL and ENDMDL for multi-model structures (such as NMR ensembles or alternate X-ray models), and CONECT records that describe explicit bonds. Metadata is stored in records such as HEADER, TITLE, COMPND, SOURCE, and AUTHOR, while CRYST1 records provide unit-cell parameters for crystallographic structures.

Limitations of the fixed-column PDB format include restricted data types and field widths, which can hinder large structures and complex metadata. This has led to the adoption of mmCIF (PDBx format) as a more flexible successor. Nevertheless, the PDB format remains in active use due to historical datasets and broad compatibility, with many programs able to read, write, and convert between PDB and other formats.