Written and maintained by CASRAI Editorial Board
Last updated
Short answer
UniChem is a free EMBL-EBI service that cross-references compound identifiers. Look up a Standard InChIKey and it returns the matching record identifiers in each source database, such as ChEMBL, ChEBI and PDBe. It maps identifiers only; it holds no property data.
UniChem (EMBL-EBI) is a free service from the European Bioinformatics Institute that cross-references compound identifiers. Give it a structure key such as a Standard InChIKey and it returns the matching record identifiers in each source database it covers, so the same molecule can be recognised in ChEMBL, ChEBI, PDBe and other resources. It maps identifiers; it does not hold property data. (It is unrelated to the pharmaceutical company of a similar name.)
Definition and scope
The 2013 Journal of Cheminformatics paper describes UniChem as a large-scale, non-redundant database of Standard InChIs with pointers to the identifiers of those structures in source databases (Chambers et al., doi:10.1186/1758-2946-5-3). The current documentation, UniChem 2.0, calls it a simple, large-scale, non-redundant database of pointers: molecules from different sources are treated as identical when they share the same Standard InChI.
Sources include primary chemistry resources such as ChEMBL and ChEBI, and resources where small molecules are a secondary concern, such as the Gene Expression Atlas and PDBe. The paper also notes external sources such as DrugBank and ZINC being added. The service was built so that structure-based links between resources can be retrieved programmatically through a REST interface.
How lookups work
The UniChem identifier (UCI)
Each new structure receives a UniChem identifier (UCI). The paper states that the association between a structure and its UCI is never updated or deleted, which makes the UCI a stable handle even when source databases change.
Standard InChI and InChIKey as the matching key
Structures are normalised using the Standard InChI (introduced with InChI software 1.02 in 2009) and its hashed form, the InChIKey. The authors flag a drawback: some distinct forms, tautomers for example, can map to the same structure, so an exact match in UniChem means an identical Standard InChI rather than proof of identical chemistry in every sense.
The compound search API
The UniChem 2.0 compound search is a POST request to https://www.ebi.ac.uk/unichem/api/v1/compounds with a JSON body. The documented fields are:
type: one ofuci,inchi,inchikeyorsourceIDcompound: the value to look up, which must matchtype(ignored when type issourceID)sourceID: used only when type issourceID
The documentation’s example body is {"type":"inchikey","compound":"LMXNVOREDXZICN-WDSOQIARSA-N"}. The response lists each matching compound with its UniChem id, InChI and standard InChIKey, plus a list of sources. For each source you get the compound ID in that source, its short and long names, a URL to the compound page, and the UniChem source ID.
Types of identifier and related concepts
UniChem works with four kinds of lookup value. The roles below come from the paper and the compound search documentation cited under Sources.
| Identifier | API type | Role in UniChem |
|---|---|---|
| UCI (UniChem identifier) | uci |
Assigned to each new structure; the association between structure and UCI is never updated or deleted. |
| Standard InChI | inchi |
The matching basis: molecules are treated as identical when they share the same Standard InChI. |
| Standard InChIKey | inchikey |
The hashed form of the Standard InChI, accepted as a search key. |
| Source compound ID | sourceID |
The identifier a source database uses for a compound; searched together with the UniChem source ID. |
A reconciliation checklist for lab and records teams
The documentation and paper describe the pieces; this checklist combines them into a working order of operations for linking records across systems. It is an editorial synthesis of the sourced behaviour above.
- Start from a structure key, not a name. Names and synonyms differ between databases; the Standard InChIKey is the key UniChem normalises on.
- Look up the InChIKey (type
inchikey) and read the returned source list for the identifier each database uses. - Store the UCI alongside the source identifiers. The structure-to-UCI link is never updated or deleted, so it is the stable anchor in your record.
- Review same-InChI collisions. If your local record distinguishes forms (for example a tautomer) that share one Standard InChI, keep that distinction in your own record, because UniChem will return them as one structure.
- Keep obsolete source IDs. UniChem never deletes source assignments; it flags them current or obsolete. The LAST_REL_U_CUR field gives the last release in which an obsolete assignment was current, though the paper cautions it does not capture the full history if an assignment became obsolete more than once. Old identifiers in your records may still resolve to the right structure, so do not purge them.
Where it matters for labs and clinics
Compound registries, electronic lab notebooks and procurement or inventory records often carry the same substance under a vendor catalogue number, a database ID and a trivial name. A cross-reference service lets a records team connect those without manual matching, and lets a researcher move from a ChEMBL record to the corresponding ChEBI or PDBe entry. For deeper property or hazard data you still go to the source databases; see the CASRAI guides to the EPA CompTox Chemicals Dashboard, DrugBank, PubChem and organic spectroscopy databases, and the overview What Is Cheminformatics?
Common misunderstandings
- It is not a property database. UniChem stores structures and pointers, not melting points, bioactivity or safety data.
- It is not the company. Unichem Laboratories is a pharmaceutical manufacturer; UniChem here is an EMBL-EBI service.
- A match is not a guarantee of identical chemistry. Matching is on Standard InChI, which can merge forms such as tautomers.
- An obsolete assignment is not a deleted one. It remains stored and flagged.
FAQ
What is UniChem?
A free EMBL-EBI service that maps compound identifiers across chemistry databases using Standard InChI as the matching key.
What is a UCI?
The UniChem identifier assigned to each new structure; its association with the structure is never updated or deleted.
Can I search by InChIKey?
Yes. The compound search accepts uci, inchi, inchikey or sourceID as the search type.
Does UniChem delete old source identifiers?
No. Assignments are flagged current or obsolete rather than deleted.
Sources
- Chambers J et al. UniChem: a unified chemical structure cross-referencing and identifier tracking system. J Cheminform 2013;5:3. PMC3616875 (read 2026-11-08).
- UniChem 2.0 documentation, EMBL-EBI. chembl.gitbook.io/unichem (read 2026-11-08).
- UniChem 2.0 compound search API. chembl.gitbook.io/unichem/api/compound-search (read 2026-11-08).
