Which databases does it search?

Around sixty registered authorities including Ensembl, UniProt, gnomAD, ClinVar, ChEMBL, PubChem, Gene Ontology and PubMed. The list comes from the connector registry rather than a short built-in selection, and the surface names the authority each record came from.

Is the data cached locally?

No. Queries go to the authority itself, so an identifier resolves against what that database says today rather than against a snapshot that has drifted. Sources that a browser cannot reach directly are documented in the owning doc.

How complete is this?

The engine is shipped and returns live records, but it is not golden-validated yet and the viewer half is still being built. The owning doc states exactly which sources were probed, which returned records and which are unreachable.

Analyze

Search Databases

Ask one question and have it answered by the database that actually owns the answer. A gene, variant, protein, compound, ontology term or paper is resolved against the live authority — Ensembl, UniProt, gnomAD, ChEMBL, Gene Ontology, PubMed and around fifty more — instead of against a local list that has quietly gone stale, and every record says which authority it came from.

Open Search Databases Open the BLAST & EMBL-EBI services sample

Nucleotide, protein, an annotated plasmid and a headerless raw sequence: reach blastx, blastp and blastn, and every EMBL-EBI service, on real queries

What it does

01One query across roughly sixty authoritative biological, chemical and literature sources
02Gene, variant, protein, compound, clinical, omics, ontology and literature lookups
03Records returned from the live authority rather than from a local copy
04Each result carries the source authority it came from
05Reachable from the editor and from the copilot, on the same engine

What it runs on

Runs on the connectors engine in packages/ports, ported from the feynman research-tools federation and recalibrated for the browser. Sixty of sixty-two registered sources return live records. Not yet golden-validated — see docs/tools/BIO_CONNECTORS-DONE.md for the measured status. The panel runs on @ports/connector-rows, @ports/connectors, @ports/manhattan.

Everything runs client-side. Files are parsed in the browser and projects are saved on this machine, so an unpublished sequence is never uploaded.

Questions

Which databases does it search?: Around sixty registered authorities including Ensembl, UniProt, gnomAD, ClinVar, ChEMBL, PubChem, Gene Ontology and PubMed. The list comes from the connector registry rather than a short built-in selection, and the surface names the authority each record came from.
Is the data cached locally?: No. Queries go to the authority itself, so an identifier resolves against what that database says today rather than against a snapshot that has drifted. Sources that a browser cannot reach directly are documented in the owning doc.
How complete is this?: The engine is shipped and returns live records, but it is not golden-validated yet and the viewer half is still being built. The owning doc states exactly which sources were probed, which returned records and which are unreachable.

Also searched as

biological database search · federated bioinformatics search · search UniProt Ensembl ChEMBL together · gene lookup tool · variant lookup gnomAD ClinVar · protein lookup UniProt · compound lookup ChEMBL PubChem · ontology lookup GO · literature search PubMed Europe PMC · query multiple bio databases at once