Example searches: iron*, InChI=1S/CH4O/c1-2/h2H,1H3, caffeine | Advanced Search

Downloads

Each month, a new version of ChEBI is published, accompanied by updated downloadable files. These files are accessible from the EBI FTP server in various formats and versions. Currently, the following files are available:

Ontology Files

ChEBI Ontology is provided in three variants: FULL, CORE, and LITE. Each variant is available in three different formats: OWL (RDF/XML serialization), OBO, and JSON (OBO Graph JSON), resulting in a total of nine different files. Tools such as Protégé, Topquadrant, OWL-API, or Pronto can be used with this ontology. Below is a brief description of each variant:

CHEBI FULL

Includes all compound information: IDs, names, definitions, subsets, relationships, synonyms, manually curated cross-references, chemical data (charge, mass, monoisotopic mass, and formula), structure data (SMILES, InChI, and InChIKey), secondary IDs, and WURCS for carbohydrates.

CHEBI CORE

A reduced version of the ChEBI ontology, containing all information included in the FULL version except for synonyms and manually curated cross-references.

CHEBI LITE

The lightest version of the ChEBI ontology which includes only ChEBI IDs, names, subsets, and relationships. This variant is suitable for those primarily interested in the modelling of relationships rather than detailed chemical information.

Note: Ontology files are updated nightly in addition to the regular monthly releases. The latest changes are available here.

Flat Files

ChEBI is stored in a relational database and we currently provide all-star ChEBI table in a flat-file tab delimited format. There are various spreadsheet tools available to import this into a relational database. The files are stored in the same structure as the relational database.

SDF Files

ChEBI provides its chemical structures and additional data in SDF format. These files are available in two variants: FULL and LITE. For each variant, there are files for all 2-star and 3-star entries, as well as separate files containing only 3-star entries. Please note that it excludes any ontological information as ontological classes are not able to be represented as they do not contain a structure.

SDF LITE

Contains four fields: molfile, chebi_id (e.g., CHEBI:15377), chebi_name (e.g., water), and star rating (e.g., 3).

SDF FULL

This is the most comprehensive version and includes all information found in the LITE variant, plus synonyms, chemical data (charge, mass, monoisotopic mass, and formula), structure data (SMILES, InChI, and InChIKey), secondary IDs, and WURCS for carbohydrates.

SQL Dumps

The ChEBI database is available as a PostgreSQL dump, including all 2-star and 3-star entries. The easiest way to import ChEBI as a relational database is by downloading the file pgsql_allstar.dump and following the instructions in the INSTALL file on the FTP server. This method is recommended for importing ChEBI.

Additionally, DDL statements are available for both schema creation and data insertion into each table. The DDL for schema creation is in pg_sql_create_tables.sql, and the folder generic_dump_all_star contains the DDL for data insertion by table. This process can be slower and requires attention to foreign key restrictions; hence, using the pgsql_allstar.dump is recommended.