uniprot-database - SKILL.md Agent Skill

name: uniprot_database description: UniProt — protein sequence, function, taxonomy, cross-references. Search proteins by query, retrieve a UniProt entry, map IDs between databases (PDB↔UniProt etc.), pull FASTA sequence, fetch metadata, run SPARQL against sparql.uniprot.org. Use whenever the user mentions a UniProt accession (e.g. P04637), asks for protein function/sequence/family info, or needs cross-DB ID mapping. Don't use for AlphaFold structures (use alphafold_database) or PDB structures (use rcsb_database). license: Unknown metadata: skill-author: VenusFactory2.

UniProt Database

Overview

UniProt KnowledgeBase (UniProtKB) is the central protein-knowledge resource. This skill exposes 6 tools spanning text search, single-entry retrieval, ID mapping, sequence/metadata download, and SPARQL queries.

Project Tools (VenusFactory2)

Tool	Args	Returns	Description
download_uniprot_search_by_query	`query` (UniProt query syntax, e.g. `organism_id:9606 AND reviewed:true`), `out_path`, optional `format`	rich JSON envelope	Search UniProtKB; results saved to file.
download_uniprot_retrieve_by_id	`uniprot_id` (e.g. `P04637`), `out_path`, optional `frmt` (`fasta` default)	rich JSON envelope	Retrieve one entry in the chosen format.
download_uniprot_mapping	`fr` (from db, e.g. `PDB`), `to` (to db, e.g. `UniProtKB_AC-ID`), `query` (comma-separated IDs), `out_path`	rich JSON envelope	Cross-database ID mapping (PDB↔UniProt, gene name↔accession, etc.).
download_uniprot_seq_by_id	`uniprot_id`, `out_path`	rich JSON envelope; FASTA file	Sequence FASTA only.
download_uniprot_meta_by_id	`uniprot_id`, `out_path`	rich JSON envelope; JSON file	Full metadata (function, taxonomy, GO, xrefs, etc.).
download_uniprot_sparql_by_query	`query` (SPARQL string), `out_dir` (required), `timeout` (default `120`s)	rich JSON envelope; SPARQL JSON file at `file_info.file_path`; `biological_metadata.row_count` + `head_vars`	Run an arbitrary SPARQL query against https://sparql.uniprot.org/sparql.

When to Use Which Tool

Goal	Tool
You have a UniProt accession, want sequence	`download_uniprot_seq_by_id`
You have a UniProt accession, want everything else	`download_uniprot_meta_by_id`
You have a gene name / PDB ID, want UniProt	`download_uniprot_mapping`
You have a free-text query, want a list	`download_uniprot_search_by_query`
You need a complex cross-resource graph query (e.g. all enzymes in pathway X catalyzing reaction Y)	`download_uniprot_sparql_by_query`

SPARQL Quick Examples

Top 5 proteins by name:

PREFIX up: <http://purl.uniprot.org/core/>
SELECT ?p ?name WHERE {
  ?p a up:Protein .
  ?p up:recommendedName / up:fullName ?name .
} LIMIT 5

Proteins in human, with a specific GO term:

PREFIX up: <http://purl.uniprot.org/core/>
PREFIX taxon: <http://purl.uniprot.org/taxonomy/>
SELECT ?protein ?name
WHERE {
  ?protein a up:Protein .
  ?protein up:organism taxon:9606 .
  ?protein up:classifiedWith <http://purl.obolibrary.org/obo/GO_0003700> .
  ?protein up:recommendedName / up:fullName ?name .
} LIMIT 20

Common Mistakes

Passing an unreviewed entry's accession when you need curated info: filter with reviewed:true in your search.
Mapping in the wrong direction: fr=PDB to=UniProtKB_AC-ID for "PDB → UniProt"; flip for the reverse.
SPARQL timeout on broad queries: add LIMIT N; default endpoint timeout is harsh. Bump timeout parameter if your query is legitimately heavy.

References

UniProt REST API
UniProt query syntax
UniProt SPARQL endpoint — interactive query editor with examples
ID mapping db list