Explorer
Reading , anonymously.
Select a table on the left to see its columns, row count and a copyable query, or open Query to run SQL.
The query runs in your browser with DuckDB-Wasm, against latest/ attached as bioc (and selected with USE bioc, so annotation.gene works as well as bioc.annotation.gene). The first run downloads DuckDB itself, a few MB. One read-only SELECT or WITH at a time, capped at 1,000 rows.
Queries that filter and use LIMIT are fast, because DuckDB fetches only the parts of each Parquet file it needs. Full scans of the big tables (raw.icite__metadata is 12 GB, annotation.icite__citation 8 GB) are slow and can run out of browser memory. For those, use DuckDB on your own machine.
Queries to try
Copy one into the box, or open it in the DuckDB Shell.
-- Every table in the release, by schema
SELECT schema_name, view_name
FROM duckdb_views()
WHERE database_name = 'bioc' AND NOT internal
ORDER BY 1, 2;-- Mouse genes by source
SELECT source, count(*) AS genes
FROM annotation.gene
WHERE taxon_id = 10090 AND valid_to IS NULL
GROUP BY source;-- The provenance of this release
SELECT release, published_at, git_sha
FROM read_json('https://data.bioconice.cancerdatasci.org/latest/manifest.json');-- One BugSigDB signature as a list of member names
SELECT bsdb_id, list(taxon_name ORDER BY member_index) AS members
FROM annotation.signature_taxon
WHERE bsdb_id = 'bsdb:83/1/1'
GROUP BY bsdb_id;