UniProt
Search curated protein records — sequence, function, features, and cross-references.
UniProt
UniProt is the standard public database of protein sequence and functional information; this connector searches it directly from inside RxCubed. It's available under Collect files. Search by protein name, gene, accession, or organism — for example “carbonic anhydrase 2 human”. See File Collection for the shared search, shortlist and save flow every connector here uses.
What to compare
The result card shows accession, protein name, gene, organism, review status, and sequence length. Reviewed records are manually curated (Swiss-Prot); unreviewed ones are automatic (TrEMBL). Prefer a reviewed record when one exists for your protein.
Everything else the record carries — function, subcellular location, features, and cross-references to other databases — sits under Details ▾.
AI scoring
UniProt is one of the two connectors that can be AI scored, alongside Literature Search. When scoring is on, each result gets a 0–1 score and a short summary, and you can add Qualities to say what matters beyond the search words — “human enzyme with a known catalytic zinc”, for example. With scoring off, results keep UniProt's own order. See Literature Search for how the score is put together; it works the same way here.
Files
- FASTA (canonical) — the sequence, selected by default. This is what sequence tools expect.
- Full record (JSON) — everything UniProt holds for the entry.
- Features (GFF) — annotated regions such as domains, binding sites, and modified residues.
Select several and each becomes its own evidence row under the same record and run.
Papers
A UniProt record cites the literature behind its annotations. Tick papers under Papers in the shortlist before saving, or open Group ▾ on a saved row afterwards, and RxCubed saves them as literature evidence in the same group. A paper already in the project is recognised rather than saved twice.
What to do next
A saved FASTA is the usual starting point for structure prediction and sequence alignment. If an experimental structure exists, collect it from RCSB PDB; if not, AlphaFold DB may have a predicted model for the same accession.
