- The Protein Data Bank (PDB) and macromolecular Crystallographic Information File (mmCIF) are two important formats used to represent and exchange three-dimensional structures of biological macromolecules. They store information about the spatial arrangement of atoms in proteins, nucleic acids, complexes, and other biological assemblies. Structural data in these formats support molecular visualization, structural bioinformatics, protein modelling, molecular docking, drug discovery, comparative analysis, and the study of molecular interactions.
- The traditional PDB format is a text-based structural format developed for the Protein Data Bank. It represents atomic coordinates alongside information about atoms, residues, chains, structural models, experimental methods, and other properties. A PDB file commonly uses the .pdb extension and contains records arranged in a defined fixed-column layout. Each line has a particular record type, such as HEADER, TITLE, COMPND, SOURCE, REMARK, ATOM, HETATM, TER, CONECT, and END. These records provide information about the structure, its biological source, experimental details, and atomic organization.
- The ATOM record describes atoms belonging to standard polymer residues, such as amino acids in proteins and nucleotides in nucleic acids. The HETATM record is generally used for non-standard residues and other non-polymer components, including water molecules, metal ions, ligands, and many cofactors. Each atom record contains fields for the atom serial number, atom name, residue name, chain identifier, residue sequence number, Cartesian coordinates, occupancy, temperature factor, and element symbol. The coordinates are usually expressed in ångströms (Å), allowing the three-dimensional positions of atoms to be examined and analysed.
- A simplified PDB atom record may look like this:
ATOM 1 N MET A 1 11.104 13.207 9.345 1.00 20.00 N
ATOM 2 CA MET A 1 12.560 13.400 9.700 1.00 18.50 C
ATOM 3 C MET A 1 13.100 14.800 9.200 1.00 19.20 C
ATOM 4 O MET A 1 12.500 15.800 9.500 1.00 21.00 O
- In these records, the atom name identifies the atom within a residue, while the residue name identifies the amino acid or other molecular component. The chain identifier distinguishes different polymer chains, and the residue sequence number identifies the residue within a chain. The X, Y, and Z values describe the atom’s position in three-dimensional space. Occupancy represents the fraction of the structural model associated with an atom at that position, while the temperature factor, also called the B-factor, provides information about atomic displacement or positional disorder as represented in the structural model. These values must be interpreted in the context of the experimental method and refinement procedure.
- The PDB format also supports information about secondary structure and molecular connectivity. Records such as HELIX and SHEET describe alpha helices and beta sheets, while SSBOND identifies disulfide bonds. CONECT records can specify connectivity between atoms, particularly for ligands and other non-standard components. MODEL and ENDMDL records allow multiple coordinate models to be stored in one file, which is particularly useful for structures determined by nuclear magnetic resonance (NMR) and for ensembles of alternative conformations. Alternate atom locations can also be represented when more than one conformation is modelled for an atom or group of atoms.
- Although the PDB format is widely supported and remains useful for many structural applications, it has important limitations. Its fixed-column structure imposes restrictions on the representation of large molecular assemblies, atom serial numbers, chain identifiers, residue numbering, and other structural information. Traditional PDB files also have limited flexibility for representing complex relationships between molecular entities, polymer sequences, experimental details, and structural annotations. These limitations become especially important for very large ribosomes, viral particles, macromolecular assemblies, and structures containing many chains or atoms. For this reason, the PDB format is no longer sufficient as the principal archival representation for all newly deposited structures.
- The macromolecular Crystallographic Information File (mmCIF), commonly called PDBx/mmCIF, was developed to provide a more flexible and extensible representation of macromolecular structures. It is based on the Crystallographic Information File (CIF) framework and uses a dictionary-based data model to describe structural information. Unlike the traditional PDB format, mmCIF is not restricted to a fixed 80-column line structure. It uses named data items grouped into categories, allowing large and complex datasets to be represented more comprehensively.
- An mmCIF file commonly uses the .cif extension, although .mmcif is also encountered. It contains data blocks, categories, and data items. A data block begins with a name such as data_1ABC, while individual data items are identified by names beginning with an underscore. Categories group related information, and their data items describe specific properties. For example, the _atom_site category contains atomic coordinate information, while categories such as _entity, _entity_poly, _struct, and _exptl describe molecular entities, polymer information, structural details, and experimental methods.
- A simplified mmCIF coordinate section may look like this:
data_example
loop_
_atom_site.group_PDB
_atom_site.id
_atom_site.type_symbol
_atom_site.label_atom_id
_atom_site.label_comp_id
_atom_site.label_asym_id
_atom_site.label_seq_id
_atom_site.Cartn_x
_atom_site.Cartn_y
_atom_site.Cartn_z
_atom_site.occupancy
_atom_site.B_iso_or_equiv
ATOM 1 N N MET A 1 11.104 13.207 9.345 1.00 20.00
ATOM 2 C CA MET A 1 12.560 13.400 9.700 1.00 18.50
ATOM 3 C C MET A 1 13.100 14.800 9.200 1.00 19.20
ATOM 4 O O MET A 1 12.500 15.800 9.500 1.00 21.00
- The _atom_site category is particularly important because it contains the atomic coordinate data required for many structural analyses. Its data items describe the atom identifier, element, atom name, component name, chain or asymmetry identifier, residue numbering, Cartesian coordinates, occupancy, and displacement parameters. The loop_ construct indicates that multiple rows of values follow the listed data-item names. This structure makes it possible to represent large numbers of atoms without relying on the fixed-column conventions of the traditional PDB format.
- One important feature of mmCIF is its ability to distinguish between different identifiers used to describe molecular entities and chains. For example, label_asym_id and auth_asym_id may represent different chain identifiers. The label identifiers are assigned within the standardized PDBx/mmCIF data model, whereas author identifiers preserve identifiers supplied by the structure’s authors. Similarly, label_seq_id and auth_seq_id can differ because standardized polymer sequence numbering may not match the residue numbering used in the original experimental structure. This distinction is important when connecting structural coordinates with biological sequences, literature, and external annotations. Software users should therefore check which identifier convention a tool expects rather than assuming that all chain and residue identifiers are interchangeable.
- The mmCIF format also provides detailed information about polymer sequences, non-polymer entities, biological assemblies, chemical components, crystallographic parameters, experimental conditions, and structural refinement. It can represent relationships between proteins, nucleic acids, ligands, cofactors, and other molecular components more systematically than the traditional PDB format. The format is also extensible, allowing additional categories and data items to be introduced through the PDBx/mmCIF dictionary without redesigning the entire file structure.
- PDB and mmCIF files may contain structures determined using different experimental methods, including X-ray crystallography, nuclear magnetic resonance spectroscopy, and cryo-electron microscopy. The files can include information about experimental resolution where applicable, refinement statistics, model quality, atomic displacement parameters, occupancies, and structural validation. However, the presence and meaning of particular fields depend on the experimental method and the available evidence. For example, crystallographic resolution has a specific meaning in X-ray diffraction and should not be treated as a universal measure of quality for every experimental structure.
- Both formats are used to represent atomic coordinates, but they do not necessarily contain identical information. Conversion from mmCIF to traditional PDB may result in loss of information because the PDB format cannot represent every feature available in mmCIF. Some identifiers may need to be shortened or remapped, and certain metadata or complex structural relationships may not be preserved. Consequently, mmCIF is generally preferable when working with newly deposited structures, large assemblies, or analyses requiring comprehensive metadata. The traditional PDB format remains useful when compatibility with older software or established workflows is required.
- Structural bioinformatics software provides several ways to read, inspect, analyse, and convert PDB and mmCIF files. Biopython is a widely used Python library that includes parsers for both formats. The PDBParser class reads traditional PDB files, while MMCIFParser reads mmCIF files. The following example demonstrates how to load a structure and inspect its chains and residues:
from Bio.PDB import PDBParser, MMCIFParser
# Read a traditional PDB file
pdb_parser = PDBParser(QUIET=True)
pdb_structure = pdb_parser.get_structure(
“protein”,
“structure.pdb”
)
# Read an mmCIF file
cif_parser = MMCIFParser(QUIET=True)
cif_structure = cif_parser.get_structure(
“protein”,
“structure.cif”
)
# Display chains and residues
for model in pdb_structure:
for chain in model:
print(“Chain:”, chain.id)
for residue in chain:
print(
residue.get_resname(),
residue.id
)
- Biopython can also be used to extract atomic coordinates, calculate distances, examine residues, and perform structural comparisons. For example, the following code calculates the distance between two selected atoms:
from Bio.PDB import PDBParser
import numpy as np
parser = PDBParser(QUIET=True)
structure = parser.get_structure(“protein”, “structure.pdb”)
model = structure[0]
chain = model[“A”]
residue_1 = chain[1]
residue_2 = chain[2]
atom_1 = residue_1[“CA”]
atom_2 = residue_2[“CA”]
distance = np.linalg.norm(
atom_1.coord – atom_2.coord
)
print(“Distance:”, distance, “angstroms”)
- Other widely used structural tools include Gemmi, which supports PDBx/mmCIF and crystallographic data processing; PyMOL, which supports structural visualization and analysis; ChimeraX, which is used for interactive molecular visualization; and the CCP4 and Phenix software suites, which provide tools for macromolecular crystallography. These tools help researchers inspect molecular structures, identify structural features, examine ligand-binding sites, analyse molecular assemblies, and prepare structures for downstream computational studies.
- PDB and mmCIF files are used in many areas of bioinformatics and computational biology. Protein structure analysis uses atomic coordinates to investigate folding, structural domains, secondary structure, and three-dimensional organization. Comparative structural analysis examines similarities and differences between proteins, including homologous proteins and proteins with related biological functions. Molecular docking and virtual screening use structural coordinates to investigate possible interactions between proteins and small molecules. Structure-based drug discovery relies on three-dimensional representations of target proteins to identify binding pockets and evaluate potential drug candidates.
- These formats are also important in protein–protein interaction studies, enzyme mechanism analysis, nucleic acid structure research, molecular dynamics preparation, and the investigation of disease-associated structural changes. Researchers can compare experimentally determined structures with predicted models, investigate mutations affecting protein stability, and study how ligands or cofactors interact with biological macromolecules. In molecular dynamics workflows, structural files are often used as starting points, although additional preparation is normally required to assign protonation states, add missing atoms, define force-field parameters, and prepare solvent and ions.
- Before using a PDB or mmCIF file, researchers should examine its structural completeness and experimental metadata. Some structures contain missing residues, unresolved loops, incomplete side chains, alternate conformations, or regions with uncertain atomic positions. A structural model may represent only part of a biological molecule or may contain multiple chains that do not correspond directly to the biological assembly. It is also important to distinguish the asymmetric unit from the biological assembly, because the deposited coordinate model may not directly represent the complete functional molecular complex. Biological assembly information should be checked before interpreting oligomeric structure or intermolecular interactions.
- The choice of file format also matters when preparing data for computational analysis. Researchers should verify the reference structure, chain identifiers, residue numbering, ligand identities, coordinate units, and experimental method. When converting between formats, they should check whether atom records, alternate locations, chain identifiers, residue numbering, ligands, and metadata have been preserved. For reproducible structural research, the PDB accession code, file format, retrieval date, structure revision or version where available, and any processing or conversion steps should be documented. Validation tools can help identify malformed records, unusual geometry, missing atoms, and other structural problems.
- The PDB and mmCIF formats form an essential part of the infrastructure of structural bioinformatics. The traditional PDB format provides a simple and widely recognized representation of molecular coordinates, while PDBx/mmCIF offers a more comprehensive and scalable framework for representing modern macromolecular structures and their associated metadata. Understanding their organization, coordinate conventions, identifiers, limitations, and software support enables researchers to retrieve, interpret, process, and analyse three-dimensional biological structures accurately and reproducibly.