Back to list

How to Use InterPro for Protein Function Analysis: From Sequence to Annotation

Published on November 8, 2026

How to Use InterPro for Protein Function Analysis: From Sequence to Annotation

Figure 1. InterPro protein function analysis concept: using human c-Src as an example, InterPro can identify SH3, SH2 and kinase domains (schematic, not to scale).


Background: Why Do We Need InterPro Protein Function Analysis?

As genomic and proteomic data grow explosively, a single amino-acid sequence rarely tells us what a protein “does”. Protein function is not determined by a single residue but by combinations of recognizable domains, families, and functional sites.

InterPro is a freely accessible resource for classifying protein sequences into families. It integrates predictive models, known as signatures, from multiple member databases to classify sequences into families and predict the presence of domains and significant sites. According to the InterPro 2025 review, it provides annotations for over 200 million sequences, ensuring extensive coverage of UniProtKB, and includes mappings to major resources such as Gene Ontology (GO), PDBe and the AlphaFold Protein Structure Database.

Key fact: InterProScan is InterPro's large-scale functional annotation tool and an essential component of the large-scale annotation pipelines of UniProt, Ensembl and MGnify Genomes.


Principle: How Does InterPro Integrate Domain Annotations?

From “signatures” to “domains”

InterPro does not “invent” function; it integrates. It aggregates predictions from multiple member databases for the same sequence:

· Pfam (protein families, domains)

· SMART (domains, repeats, signal peptides)

· PROSITE (patterns and profiles, including functional sites)

· PRINTS (fingerprints)

· Gene3D (CATH-Gene3D, structure-based evolution)

· PANTHER (families and pathways)

· CDD / SSF (Conserved Domain Database, SCOP superfamilies)

· TIGRFAMs, PIRSR, AntiFam (families, conserved sites, anti-predictors)

When a sequence matches signatures from several member databases, InterPro's post-processing filters, integrates and compares them into a single integrated InterPro entry, and maps the result to GO terms and pathways.

Taking human proto-oncogene tyrosine-protein kinase c-Src (UniProt P12931, 536 aa) as an example, the InterPro query returns 36 records: 18 integrated and 7 unintegrated entries, covering several InterPro families/domains, and mapped to protein tyrosine kinase activity (GO:0004713), ATP binding (GO:0005524), protein phosphorylation (GO:0006468), protein binding (GO:0005515), and protein kinase activity (GO:0004672).

 

Figure 2. c-Src domains and functional sites InterPro can distinguish the SH2 and SH3 domains, the tyrosine kinase (PTK) catalytic domain and its activation loop

Figure 2. c-Src domains and functional sites: InterPro can distinguish the SH2 and SH3 domains, the tyrosine kinase (PTK) catalytic domain and its activation loop (schematic, not to scale).


Layered annotation: family / domain / site

InterPro annotations are layered, which matters for downstream engineering:

Level

Meaning

Example (c-Src)

Family

Superfamily/family membership

Protein tyrosine kinase family

Domain

Independently folding/evolving module

SH2, SH3, kinase (PTK) domain

Repeat

Tandem repeat units

Ankyrin repeats in some proteins

Site

Specific catalytic/binding residues

Kinase ATP-binding site, activation loop

Note: Figures 1 and 2 use a real structure (AlphaFold model AF-P12931-F1) rendered with PyMOL and colored by InterPro domains; annotated as schematic and not to scale.


Measurement: How to Choose Member Databases and Annotation Types?

Choosing which type of InterPro annotation to use depends on your question:

· “What protein is this?”: focus on family/superfamily (PANTHER, Pfam, Gene3D) to establish identity.

· “Where are the functional modules?”: look at domains (Pfam, SMART, CDD, SSF) for module recombination and site design.

· “Which are the key residues?”: look at functional sites/patterns (PROSITE, PRINTS, PIRSR) for point mutation and active-site engineering.

· “Exclude false positives”: use anti-predictors such as AntiFam to remove spurious matches.

Note: query_interpro_meta_by_uniprot_id returns database annotations for already-curated proteins and does not run InterProScan; for a novel sequence, run InterProScan yourself (below).


Conditions: Running InterProScan and Interpreting Results

How to run

InterProScan supports scanning protein or nucleotide sequences (nucleotide requires the -t option). For local large-scale workflows, InterProScan 6 is recommended — it is a complete reimplementation as a Nextflow pipeline that improves scalability, portability and reproducibility across local, HPC and cloud environments, and returns pre-computed annotations in JSON via the redesigned Matches API.

Note: InterProScan 5.78–109.0 is the final release of version 5; upgrade to InterProScan 6 to annotate sequences with future InterPro releases.

Key parameters (schematic)

· Input: FASTA protein sequence (or nucleotide, with -t)

· Applications: choose member databases by annotation type (HMMER, Profile, Pattern, etc.)

· Output: TSV / XML / JSON (InterProScan 6 Matches API is JSON)

· Storage: EBI results retained for at least 24 hours

 

Figure 3. InterProScan 6 large-scale functional annotation workflow (schematic)

Figure 3. InterProScan 6 large-scale functional annotation workflow (schematic).


How to interpret

· Prioritize integrated entries — they are more reliable than hits from a single member database.

· Verify the source database (Pfam/SMART/PROSITE…) to know whether it is a family, domain or site annotation.

· Use GO terms to validate functional semantics, avoiding function misjudgment from domain names alone.

· For key conclusions (e.g., active sites), confirm with experimental or structural evidence — InterPro provides predictive annotation, not experimental evidence.


How MatwingsVenus® Empowers InterPro Protein Function Analysis

InterPro provides reliable sequence annotation, but turning it into wet-lab-validated candidates and an AI closed-loop still requires engineering capability. MatwingsVenus® embeds InterPro protein function analysis into its dry-wet closed-loop platform:

· Wet Lab Center: design protein-engineering and active-site candidates based on InterPro domain/site annotations.

· Expert Center: domain experts cross-validate InterPro annotations and downstream experimental plans and choices.

· Dry-Wet Closed-Loop Platform: integrate InterPro family/domain/site information into the protein-design and directed-evolution AI loop for “annotate → design → validate → feedback” iteration.

In this way, InterPro predicts, MatwingsVenus® validates and optimizes — together forming a complete path from sequence function analysis to protein engineering.


FAQs

Does InterPro protein function analysis directly tell me the function?

InterPro provides predictive family/domain/site annotations based on member-database signatures. It is very helpful for inferring function, but it is prediction, not experimental fact. Key conclusions (especially active sites) should be confirmed with experimental or structural evidence.

What is the relationship between InterPro and InterProScan?

InterPro is the database/annotation resource, while InterProScan is its large-scale scanning tool. InterPro stores annotations for curated, known proteins; InterProScan runs signature searches on user-submitted new sequences and generates annotations. For a novel sequence, run InterProScan.

Which member database annotation should I choose?

It depends on the question: use family/superfamily (PANTHER, Pfam, Gene3D) for identity; use domains (Pfam, SMART, CDD, SSF) for functional modules; use functional sites (PROSITE, PRINTS, PIRSR) for key residues; use AntiFam to exclude false positives.

InterProScan is slow — how can I speed it up?

InterProScan is computationally expensive, sometimes taking minutes for a single sequence. InterProScan 6 improves scalability and data management; additionally, using pre-computed matches (Matches API) via checksum lookup reuses existing annotations and avoids recomputation.

What is the relationship between InterPro annotations and GO terms?

Integrated InterPro entries are mapped to Gene Ontology (GO) terms (molecular function, biological process) to unify functional semantics. GO enables further pathway/ontology-level analyses — for example, c-Src is mapped to “protein tyrosine kinase activity” and “protein phosphorylation”.

Can InterPro be used to design or engineer proteins?

InterPro mainly answers “what is this and which modules/sites does it have”, so it is best for pre-design functional positioning. To actually engineer candidates, combine it with protein-design workflows (mutation-site planning, stability/developability) and AI-directed evolution.

Can I conclude from InterPro alone without experimental data?

You can use it for preliminary functional positioning and hypothesis generation, but InterPro is predictive. Without experimental or structural evidence, it should not serve as the final functional conclusion, especially for active sites and clinically relevant inferences. Distinguish 【prediction】 from 【measured/validated】.


Summary

InterPro protein function analysis integrates signatures from multiple member databases to provide layered annotations of families, domains and functional sites, and maps them to GO and other ontologies. It is an authoritative entry point for sequence function annotation. Understanding its integration logic, selection principles, and InterProScan operating conditions significantly improves the credibility of functional inference.

Building on this, MatwingsVenus® embeds InterPro annotations into wet-lab validation and AI protein-design loops, advancing from “prediction” to “validation and optimization” — a complete path from sequence function analysis to protein engineering. If you are advancing protein function annotation, directed evolution, or AI protein design, explore MatwingsVenus's Wet Lab Center, Expert Center, and Dry-Wet Closed-Loop Platform.

Scan to join community

Scan to join community