简体中文
English
Website review
This review may include AI-generated content and is not individually verified. Domain, certificate and other facts come from the observed data on this page.
What is UniProt?
UniProt is a free, comprehensive reference resource for protein sequence and functional information. It is widely used in life sciences research and is known as a leading public protein knowledgebase, offering curated data on proteins across many organisms.
What it provides
- Protein sequences and related functional annotations.
- Links to other biological data resources (for example, gene, structure, and pathway databases).
- A stable, citable reference for protein information used in research and analysis.
Who uses it
- Molecular biologists, biochemists, and bioinformaticians.
- Students and educators learning about proteins and genomics.
- Researchers who need to look up a specific protein, compare sequences, or gather functional details.
A practical way to start
If you already have a protein name, gene name, or accession number, go to UniProt and use the search box. From there, open the relevant entry to see sequence, function, taxonomy, and cross-references. If you are browsing without a specific target, use the advanced search filters to narrow by organism, protein family, or reviewed status.
A useful decision criterion: choose UniProt when you need a broad, curated protein reference with rich annotation and cross-links. If you need only a quick sequence lookup or a single database record, UniProt may be more than you need; if you need to integrate protein data with other biological datasets, its cross-references make it a strong starting point.
How can I search for a specific protein sequence in UniProt?
To search for a specific protein sequence in UniProt, use the site's sequence search rather than a plain keyword search. A protein name or gene keyword can return many related entries; pasting the actual sequence looks for matching or similar sequences, which is what you want when you already have the residues.
On UniProt, the practical route is:
- Go to the search bar and choose the sequence search option (often labeled as a peptide or sequence search) instead of the standard text query.
- Paste your sequence in single-letter amino acid code, without spaces, line numbers, or FASTA header lines. If you paste a FASTA record, keep only the sequence portion.
- Choose the database scope. Searching UniProtKB returns reviewed and unreviewed entries; restricting to reviewed (Swiss-Prot) gives curated records but may miss less-studied proteins.
- Run the search and review the results list, which typically ranks exact or high-identity matches first.
A concrete scenario: you have a short peptide from a mass-spectrometry experiment and want to know which protein it belongs to. A text search for a guessed protein name is unreliable here. A sequence search against UniProtKB, then narrowing to reviewed entries, quickly shows whether the peptide maps to a known protein or only to a predicted one.
Two trade-offs matter:
| Choice | What you get | What you give up |
|---|---|---|
| Reviewed only | Curated, well-annotated entries | Less coverage of novel or poorly studied proteins |
| All UniProtKB | Broader coverage | More low-confidence or automatically annotated hits |
If your sequence is very short, expect many incidental matches; if it is a full-length protein, an exact match usually points to a single entry. For a sequence from a non-model organism, start broad, then filter by organism or taxonomy if the result list is long.
Next step: if you get no convincing match, try the same sequence against a related sequence database such as NCBI, since UniProt may not include every deposited sequence.
What tools does UniProt offer for sequence alignment and similarity searches?
UniProt's alignment and similarity-search tools are reached through its web interface rather than as a separate downloadable suite, and the page you land on may require JavaScript and a current browser to render them. The fallback message on the site itself notes that a blank page usually means JavaScript is disabled, the browser is outdated, or the network connection dropped — so if the tools seem missing, enable JavaScript and reload first.
What you can actually do
- BLAST-style similarity search. You paste a protein sequence (or a UniProt accession) and search it against UniProt's sequence collections to find related proteins, then inspect the hits' annotations.
- Multiple sequence alignment. Align several sequences to see conserved regions, gaps and residue-level differences across a family or set of homologues.
- Pairwise alignment. Compare two sequences directly when you only need to judge how closely they match.
- Cross-references to dedicated tools. UniProt entries link out to specialist alignment and structure resources, so you can move from a search hit into a deeper analysis without re-entering the sequence.
Who benefits most
A wet-lab researcher checking whether an uncharacterised protein resembles a known enzyme gets the most value from the similarity search plus the annotation columns that come with each hit. A bioinformatician building a family alignment will more often export sequences and run the alignment in a purpose-built tool, using UniProt mainly as the sequence and annotation source.
A practical starting point
Take one sequence of interest, run a similarity search against UniProt, and sort the results by the significance measure the tool reports rather than by the first hit listed. Then open two or three top hits and compare their functional annotations — if they disagree, that disagreement is usually the interesting part of the analysis.
Related resources worth knowing
- UniProt — the sequence and functional annotation hub itself.
- NCBI — home of the BLAST family of search programs, useful when you want more control over scoring and database choice.
- EMBL-EBI — hosts several alignment and similarity services that complement UniProt searches.
- InterPro — for classifying a protein into families and domains once you have a candidate sequence.
Decision criterion
Use UniProt's built-in search when you want similarity results tied directly to curated functional annotation. Switch to a dedicated alignment platform when you need fine control over gap penalties, substitution matrices, or large batch alignments.
How do I download protein data from UniProt for use in my research?
UniProt is a public protein sequence and functional annotation resource, and downloading its data is a normal part of a bioinformatics workflow. The exact steps depend on whether you need a handful of sequences or the full database.
Quick downloads for a few proteins
If you already know which proteins you want, use the search interface and export results directly.
- Search by gene name, accession number, organism or keyword.
- Select the entries you need, then choose a download format.
- Common formats: FASTA (for sequence analysis), tab-separated or Excel (for annotation tables), XML/JSON (for programmatic parsing), GFF (for genomic features).
FASTA is the default choice for BLAST, alignment or phylogenetics. Spreadsheet formats suit literature review, annotation comparison or building a candidate list.
Bulk downloads for whole proteomes or the full database
For large-scale work, use the download section rather than the search page. This is the right route when you need:
- All reviewed (Swiss-Prot) entries — manually curated, higher confidence, smaller.
- All unreviewed (TrEMBL) entries — broader coverage, automatically annotated, much larger.
- A complete proteome for one organism.
- Specific releases for reproducibility, so your analysis can be repeated exactly.
Reviewed entries are the safer starting point for functional claims; unreviewed entries are useful for coverage but need extra validation. This trade-off matters most in comparative genomics and machine-learning datasets, where annotation quality directly affects results.
A practical decision path
| Your situation | Recommended route |
|---|---|
| A few known proteins | Search and export FASTA or TSV |
| One organism's proteome | Organism-specific download |
| High-confidence annotations only | Reviewed (Swiss-Prot) set |
| Maximum coverage | Reviewed plus unreviewed |
| Reproducible pipeline | Pin a specific release version |
One thing to check first
If the site appears blank, the cause is often browser-side. Enable JavaScript, update your browser, and confirm your network connection is stable before reloading. This is the situation described in UniProt's own fallback message on UniProt.
As a next step, decide whether your analysis needs curated annotations or maximum sequence coverage, pick the matching dataset, and record the release version in your methods section.
What is the difference between UniProtKB/Swiss-Prot and UniProtKB/TrEMBL?
UniProtKB is the protein knowledgebase, and it is split into two sections: Swiss-Prot, the manually reviewed part, and TrEMBL, the unreviewed part that is enriched automatically. The core difference is human curation versus computational annotation, which affects trust, detail and coverage.
How they differ in practice
| Aspect | UniProtKB/Swiss-Prot | UniProtKB/TrEMBL |
|---|---|---|
| Curation | Manually reviewed by curators | Automated annotation, not reviewed |
| Content | Higher confidence; entries are consolidated and checked | Generated from translation of nucleotide sequences; may contain errors or redundancy |
| Coverage | Smaller, slower to grow | Much larger, grows quickly with new sequencing data |
| Best for | Publication-level work, clinical or structural studies, training data where accuracy matters | Exploratory searches, finding homologs, broad surveys where recall matters more than precision |
Which one to use
- Use Swiss-Prot when you need reliable functional annotation, clean identifiers, or a curated reference set for an analysis.
- Use TrEMBL when you need maximum coverage — for example, searching for distant homologs or checking whether any sequence similar to your protein has been recorded.
- A common approach is to search both, then filter or prioritise reviewed entries, treating unreviewed hits as hypotheses to verify.
A concrete example
If you are annotating a newly sequenced bacterial genome, you might first search TrEMBL to capture every possible match, then check whether the key hits also appear in Swiss-Prot. A TrEMBL-only hit is worth investigating but should be labelled as unreviewed in your results.
Next step
Start at UniProt and use the reviewed/unreviewed filter in the search results; if a protein matters to your conclusion, open its entry and check the review status before citing it.
How can I access UniProt programmatically via its API or FTP site?
UniProt offers both an API and FTP access for programmatic use. The API is best for querying specific proteins, sequences, or annotations on demand; FTP is better for downloading entire datasets in bulk for local analysis.
API access
UniProt's REST API lets you retrieve entries, search by keywords, and fetch specific fields. Typical uses include:
- Fetching a protein record by accession (e.g.,
P12345) - Running a search query and paginating through results
- Requesting only the fields you need (sequence, function, cross-references) to keep responses small
Use the API when your workflow is interactive or selective. Expect rate limits and handle HTTP errors and retries in your code.
FTP access
The FTP site hosts downloadable releases of UniProt databases, including:
- Reviewed (Swiss-Prot) and unreviewed (TrEMBL) sets
- Sequence and annotation files in standard formats
- Supporting files such as cross-reference mappings
Use FTP when you need a complete, versioned snapshot for pipelines, local BLAST databases, or reproducible analyses. Downloads are large, so plan storage and bandwidth.
Quick comparison
| Need | Better choice |
|---|---|
| One protein or a small set of records | API |
| Full database or large subset | FTP |
| Frequent, changing queries | API |
| Reproducible, versioned dataset | FTP |
Practical next step
If you are building a script, start with the API for a single accession to confirm your parsing works, then switch to FTP for the full dataset you need. For official documentation and download locations, see UniProt.
GoDaddy
Cambridge, Cambridgeshire, United Kingdom
User reviews (0)