Written and maintained by CASRAI Editorial Board
Last updated
InterPro is a freely accessible resource, run from EMBL-EBI, that classifies protein sequences into families and predicts domains and important sites by combining the predictive models of several member databases. Researchers use it to annotate a protein of interest, to look up which families and domains a sequence belongs to, and to run the same classification locally on their own sequences with InterProScan. This page maps what InterPro contains, how its member databases fit together (including Pfam and SMART), how to reach the data, and how to cite it. Facts below come from the InterPro documentation and the 2025 database paper in Nucleic Acids Research, checked on 2026-11-07.
What InterPro is
The InterPro paper describes it as “a freely accessible resource for the classification of protein sequences into families”. It integrates predictive models, called signatures, from its member databases into single InterPro entries, and provides annotations for more than 200 million sequences (Blum et al., Nucleic Acids Res 53(D1):D444-D456, 2025, doi:10.1093/nar/gkae1082, covering release 101.0).
Two ideas are worth separating. A signature is a model contributed by a member database. An InterPro entry groups signatures that describe the same family, domain or site, so one entry can stand for several member-database models. In release 101.0, the paper reports 84,588 signatures, of which 69% were integrated into 45,899 InterPro entries. The remaining signatures are present in their member databases but not yet merged into an entry.
Who runs it and what it covers
The database paper’s lead authors are at EMBL-EBI (European Bioinformatics Institute, Hinxton, UK), with co-authors at other institutions including NCBI, the Swiss Institute of Bioinformatics and the University of Padova. The documentation site lists an “InterPro team” and a “Funding” section.
Member databases
The documentation lists 13 member databases: CATH-Gene3D, CDD, HAMAP, MobiDB Lite, NCBIfam, PANTHER, Pfam, PIRSF, PRINTS, PROSITE profiles, SFLD, SMART and SUPERFAMILY. It lists AntiFam separately. The 2025 paper names 12 member databases in its introduction, and its Table 1 splits PROSITE into patterns and profiles. Counts therefore differ by source and release; check the documentation page for the current list.
Pfam and SMART
Both are member databases of InterPro. The Pfam change matters for anyone who still bookmarks the old site: the paper states that the standalone Pfam website was decommissioned and “Pfam is now exclusively accessible through InterPro”, with automatic redirection beginning in October 2022. SMART is listed among the member databases that contribute signatures to InterPro.
Coverage and releases
In release 101.0, at least one InterPro match was found for 81.8% of unreviewed and 96.8% of reviewed UniProtKB sequences. New InterPro releases appear every eight weeks, one day after each UniProt release. For what the matched sequences are, see the CASRAI guide to the UniProt protein sequence and functional annotation knowledgebase.
How to access InterPro data
The InterPro paper and the documentation describe these routes.
- Website: https://www.ebi.ac.uk/interpro, the browsable interface.
- Download and Export: the documentation has a “How to download InterPro data?” section covering the Download page, the Export button and “Your downloads”.
- FTP: bulk files at https://ftp.ebi.ac.uk/pub/databases/interpro/.
- API: the InterPro Application Programming Interface, at https://www.ebi.ac.uk/interpro/api/.
InterProScan
InterProScan “takes protein or nucleic acid sequences and searches them against InterPro’s predictive models”. It is the route for classifying your own sequences, for example a newly assembled proteome, rather than looking up existing entries. The paper states it is distributed under the open source Apache 2.0 licence. The documentation lists InterProScan sections for web services, web based tools, source code and previous releases. Release notes pair the two version numbers, for example “InterPro 84.0 and InterProScan 5.50-84.0”, so record both the InterPro release and the InterProScan version you used.
How to cite InterPro
The documentation has a “Citing InterPro” section with the latest publications and all previous publications. A sensible practice is to cite the current database paper for the resource and to state in your methods the InterPro release number and, if you ran it, the InterProScan version, because signatures and entries change between releases and the same sequence can return different matches.
Strengths and limits
- Strength: one query returns evidence from many complementary methods, so a domain call can be cross-checked across member databases.
- Strength: InterProScan is open source (Apache 2.0), and data are available through the web, bulk FTP and an API.
- Limit: 31% of release 101.0 signatures were not integrated into InterPro entries, and 18.2% of unreviewed UniProtKB sequences had no match.
- Limit: the paper reports 3,887 AI-generated entries added between releases 99.0 and 101.0, and a deep-learning annotation (Pfam-N) giving at least one annotation to 85% of sequences in UniProtKB 2024_04. Treat computationally generated annotation as prediction, not experimental evidence.
- Limit: figures here are for release 101.0; later releases will differ.
Alternatives and related resources
| Resource | Relationship to InterPro |
|---|---|
| Pfam | Member database; reached through InterPro since the standalone site was decommissioned |
| SMART | Member database contributing signatures |
| CDD | Member database |
| UniProtKB | Sequence set InterPro annotates; InterPro releases follow UniProt’s schedule |
For predicted structures of the same proteins, see the AlphaFold Protein Structure Database guide; for computing physical properties of a sequence, see ProtParam.
FAQ
What is InterPro used for?
Classifying protein sequences into families and predicting domains and sites by combining member-database models. Use the website to look up proteins or entries and InterProScan to classify your own sequences.
Is InterPro the same as Pfam?
No. Pfam is one of InterPro’s member databases, and it is now accessed only through InterPro.
Where do I download InterPro data?
From the website’s Download page and Export button, in bulk from the FTP site, or through the API, as listed above.
Is InterPro free?
The paper describes it as freely accessible, and InterProScan is distributed under the Apache 2.0 licence.
How often does InterPro update?
Every eight weeks, one day after each UniProt release.
How do I cite InterPro?
Cite the current InterPro database paper (Blum et al., 2025, doi:10.1093/nar/gkae1082) and state the InterPro release and, if used, the InterProScan version.
Sources
- InterPro documentation, https://interpro-documentation.readthedocs.io/en/latest/, checked 2026-11-07.
- Blum M, Andreeva A, et al. InterPro: the protein sequence classification resource in 2025. Nucleic Acids Res 53(D1):D444-D456, doi:10.1093/nar/gkae1082, https://academic.oup.com/nar/article/53/D1/D444/7943356, checked 2026-11-07.
- InterPro website and FTP: https://www.ebi.ac.uk/interpro and https://ftp.ebi.ac.uk/pub/databases/interpro/, checked 2026-11-07.
