UniProt Search FAQ

These answers cover what UniProt contains, how to search by gene name versus accession ID, how to retrieve the full protein sequence, the difference between Swiss-Prot and TrEMBL entries, chaining results into downstream tools, and how current the data is.

Frequently asked questions

What is UniProt and what databases does it include?

UniProt (Universal Protein Resource) is the most comprehensive protein knowledge base, containing over 250 million protein entries. It is maintained by EMBL-EBI, SIB, and PIR. UniProt includes two main sections: Swiss-Prot (~570,000 manually curated entries with high-quality annotations reviewed by expert biologists) and TrEMBL (~250 million computationally annotated entries generated from genome sequencing projects). UniProt provides protein function, disease associations, Gene Ontology terms, domains, post-translational modifications, full amino acid sequences, and cross-references to PDB, ChEMBL, AlphaFold, and many other databases.

How do I search by gene name vs. accession ID?

Give the accession for a direct entry lookup when you have a specific UniProt ID (e.g., P00533). This is the fastest method and returns the exact protein record. Give a gene symbol for a gene search (e.g., EGFR, TP53). Gene searches may return multiple entries if the symbol maps to multiple proteins or is present in multiple organisms. Name the organism (the search covers all organisms unless you do) to disambiguate. You can also describe the protein in free text to search by protein name, function, or disease.

Can I get the full protein sequence?

Yes. Ask for the sequence in your question. By default, sequence retrieval is disabled to keep responses compact and fast. When enabled, the full amino acid sequence is included in the response and can be chained directly into ESMFold for structure prediction, BLAST for similarity search, or ESM-3 for embedding generation.

What is the difference between Swiss-Prot and TrEMBL entries?

Swiss-Prot entries are manually curated by expert biologists who review experimental evidence from published literature. They have comprehensive functional annotations, disease associations, and validated cross-references. TrEMBL entries are computationally annotated from genome sequencing data using automated pipelines. TrEMBL annotations may be incomplete or contain prediction errors. Always check the entry_type field in results to assess annotation reliability.

Can I chain UniProt results into other tools?

Yes. UniProt is a central hub in the platform workflow ecosystem. Retrieved sequences chain into ESMFold (AI structure prediction), BLAST (sequence similarity search), and ESM-3 (protein embeddings). Cross-references in UniProt results link directly to PDB structures for experimental structure retrieval, ChEMBL for bioactivity data, and AlphaFold for pre-computed structure predictions. The orchestrator handles chaining automatically when you request follow-up actions.

Why does my search return TrEMBL entries instead of Swiss-Prot?

Swiss-Prot coverage is limited to approximately 570,000 curated entries, focused on well-studied proteins from model organisms. If your protein of interest does not have a Swiss-Prot entry, the search returns TrEMBL (unreviewed) entries instead. This is common for proteins from non-model organisms or recently discovered proteins. TrEMBL annotations are less reliable but still useful for basic information. Check the entry_type field in results to identify which type of entry you received.

How current is the UniProt data?

UniProt releases monthly major updates with continuous incremental updates between releases. New protein entries from genome sequencing projects appear in TrEMBL first, then may be promoted to Swiss-Prot after expert review, which can take months or years. Cross-references to PDB, ChEMBL, AlphaFold, and other databases are also updated with each monthly release. The data accessed through the platform always reflects the current state of the UniProt REST API.

FAQ

Read next