The Crystallography Open Database (COD) is a free, open-access repository of experimentally determined crystal structures — primarily small organic molecules, metal–organic compounds, and inorganic materials with small to medium-sized unit cells. Founded in 2003 and maintained by a research center at Vilnius University, COD collects structures in the Crystallographic Information File (CIF) format defined by the International Union of Crystallography (IUCr) and makes every entry freely downloadable, reusable, and re-distributable without restriction. As of late 2024 it held more than 520,000 entries, making it one of the largest open collections of crystal structure data in any scientific field.
For research data managers, COD is a useful case study in domain-repository design: it is community-curated rather than commercially licensed, it accepts both published and unpublished (“personal communication” or pre-publication) depositions directly from researchers, and it runs its data through automated syntax and completeness checks before making entries public. This guide covers what COD contains, how it differs from other structural databases researchers commonly encounter, how deposition and retrieval work, and where it fits in a data management plan for crystallography or structural chemistry research.
What Counts as an Instance of COD Data
An entry qualifies for COD if it is a determined crystal structure — typically from single-crystal or powder X-ray diffraction, though other diffraction techniques are represented — expressed as a CIF file with the atomic coordinates, unit cell parameters, space group, and associated experimental metadata needed to reconstruct the structure. COD’s scope is deliberately broad within “small molecule” crystallography: organic compounds, metal–organic frameworks and coordination complexes, and inorganic structures with small to medium unit cells are all in scope. Large biological macromolecules (proteins, nucleic acids) are out of scope for COD; those belong in the Protein Data Bank (PDB) instead.
Founding, Governance, and Open-Access Model
COD began in February 2003, prompted by a call within the powder-diffraction research community (the Structure Determination by Powder Diffractometry mailing list) for a freely accessible alternative to the fee-based commercial crystallographic databases already in use. It is maintained by a team based at Vilnius University, with the project’s foundational description published as Gražulis et al., “Crystallography Open Database (COD): an open-access collection of crystal structures and platform for world-wide collaboration,” Nucleic Acids Research 40, D1 (2012). All COD data is released for unrestricted reuse — structures, derived results, and re-dissemination are all permitted without a licensing fee or use restriction, which is the core distinction between COD and the commercial databases covering overlapping structural chemistry ground. The database is mirrored at multiple sites worldwide and offers public SQL query access for bulk or programmatic retrieval, in addition to its standard web search interface.
How COD Differs from Other Crystallography and Chemistry Databases
Researchers new to structural data management sometimes conflate COD with other repositories that sound similar but serve different scopes. The distinctions matter for choosing where to deposit or search:
- COD vs. Cambridge Structural Database (CSD) — The CSD, maintained by the Cambridge Crystallographic Data Centre, covers overlapping small-molecule and metal–organic structural chemistry but is a subscription/licensed product with its own curation pipeline. COD covers similar structural territory but is entirely open-access and community-deposited, with no license required to access or reuse the data.
- COD vs. Inorganic Crystal Structure Database (ICSD) — ICSD, jointly operated by FIZ Karlsruhe and the U.S. National Institute of Standards and Technology (NIST), specializes in fully inorganic crystal structures and is also a licensed, subscription-access product. COD’s inorganic coverage overlaps ICSD’s scope but again without the access fee.
- COD vs. Protein Data Bank (PDB) — PDB is the global archive for macromolecular structures (proteins, nucleic acids, and large complexes), typically solved by X-ray crystallography, NMR, or cryo-EM. COD’s “small to medium-sized unit cell” scope is explicitly the complementary territory: small organic and inorganic molecules, not biological macromolecules. See CASRAI’s PDB guide for the macromolecular side of structural data management.
- COD vs. PubChem — PubChem, operated by NIH/NLM’s National Center for Biotechnology Information, is a compound-and-bioassay repository: it stores computed and curated 2D/3D chemical structures alongside biological assay results, and much of its content is aggregated or predicted rather than experimentally determined by diffraction. COD, by contrast, holds only experimentally solved crystal structures with the raw crystallographic parameters (unit cell, space group, atomic coordinates) that a diffraction experiment actually produces. A compound can reasonably appear in both: PubChem as a chemical/bioassay record, COD as the diffraction-derived structural determination of that same compound. They answer different questions — “what is this compound and what has it been tested against” (PubChem) versus “what is its experimentally determined three-dimensional crystal structure” (COD) — and neither substitutes for the other in a data management plan that needs both angles covered.
Data Format and Structure: CIF
Every COD entry is a Crystallographic Information File (CIF), the standard machine-readable format defined and maintained by the IUCr for representing crystallographic data as labeled key-value tags — unit cell dimensions, space group, fractional atomic coordinates, thermal displacement parameters, and citation/experimental metadata. Because CIF is a single, widely adopted standard across the crystallography community (used by the CSD, ICSD, and most journal supporting-information submissions), a structure deposited to COD is immediately interoperable with the software crystallographers already use for structure visualization, refinement checking, and further analysis, without a format-conversion step.
Deposition and Quality Control
Researchers can deposit structures directly to COD, including structures that have not yet been published elsewhere, as personal communications or pre-publication depositions — a distinguishing feature relative to databases that only accept structures tied to a completed journal publication. Submitted CIFs go through automated checks for syntactic correctness and completeness, and COD maintainers perform manual review where automated checks flag possible semantic errors (for example, internally inconsistent unit cell or coordinate data), before an entry is made public. This checked-but-open model is part of why COD functions well as a companion resource for supporting information: a journal article’s crystallographic supporting information can be deposited to COD as an independently citable, openly accessible record, distinct from and complementary to the journal’s own repository.
Retrieval, Reuse, and Downstream Applications
COD structures can be searched via its web interface (by chemical formula, unit cell parameters, space group, author, or text search) or queried directly via public SQL access for bulk retrieval and programmatic workflows, such as training data sets for crystal-structure-prediction or machine-learning models, or systematic surveys across large numbers of related structures. Because every structure is released without reuse restriction, downstream applications extend beyond conventional visualization and analysis software: COD-derived 3D structural data has also been used as source input for 3D-printing physical molecular models, an application made feasible specifically because the underlying atomic coordinate data carries no licensing barrier.
Where COD Fits in a Data Management Plan
For a crystallography or structural-chemistry research project, COD is worth naming explicitly in a data management plan (DMP) as the designated repository for crystal structure outputs, alongside (not instead of) a general-purpose or institutional repository for the rest of a project’s research data. Because COD is domain-specific, discipline-appropriate, and free to deposit to, it typically satisfies funder expectations that data be deposited in “the most appropriate disciplinary repository” rather than a generic catalog-all repository, a distinction increasingly emphasized in funder FAIR data guidance. Structures deposited to COD receive a stable identifier and URL, supporting the kind of persistent, citable data record that funder and journal data-availability policies increasingly require.
Frequently Asked Questions
Is the Crystallography Open Database peer-reviewed?
Not in the journal sense. COD applies automated syntax/completeness checks and manual review for likely semantic errors, but structures can be deposited independently of, or prior to, journal publication. Many entries do correspond to structures published and peer-reviewed elsewhere (with the CIF deposited as supporting information), but COD accepts pre-publication and personal-communication depositions as well, so presence in COD is not itself a peer-review signal.
Does COD charge for access or impose reuse restrictions?
No. COD data is released as open access with no license fee and no reuse restriction — structures, and results derived from them, may be freely downloaded, used, and re-disseminated. This is COD’s central differentiator from fee-based structural databases covering similar chemical territory.
Can I deposit unpublished crystal structures to COD?
Yes. COD’s website supports registered users depositing both published and as-yet-unpublished structures, submitted as personal communications or pre-publication depositions.
What file format does COD use, and is it interoperable with other crystallography software?
COD stores structures as Crystallographic Information Files (CIF), the IUCr-defined standard format used broadly across crystallographic software, journal supporting-information workflows, and other major structural databases, so COD entries are directly usable without conversion in the tools crystallographers already use.
How is COD different from depositing crystallographic supporting information with a journal?
Journal supporting information is typically attached to, and only discoverable through, the specific article. Depositing the same CIF to COD gives the structure an independent, openly accessible, and separately citable record that remains retrievable and searchable even if the original article’s supplementary files become hard to access, and makes the structure discoverable through COD’s own search and bulk-query tools rather than only through the journal.
