Change log

R2.0 is the final release of the v2 data, the version described in the BGC Atlas paper. The data is unchanged from R2.0-beta.6: 16,109,198 BGCs, 629,872 listed families. Family names drop the beta tag (GCF-v2-000042); the R2.0-beta.6 names (GCF-v2b6-…) and the pet names still lead to the same families. Names of the families of earlier clusterings and the BGC ids withdrawn on 5 October no longer resolve. The beta releases are retired; their notes and checksums remain on the releases page.

The gene cluster families were corrected and recomputed on the same BGCs (data release R2.0-beta.6): 629,872 listed families, 88.3% of BGCs in a listed family. Family names carry the new tag (GCF-v2b6-000042). Names of the 13 October families (GCF-v2b5-…) no longer resolve; links to earlier families still show where their members went.

124,176 BGCs added: 66,549 from 173 metalog samples, and two long-read sets, Great Barrier Reef seawater (38,505) and the Helgoland 2020 spring bloom (19,122). The families were recomputed on all of them (data release R2.0-beta.5): 618,695 listed families, 88.4% of BGCs in one. Family names carry the new tag (GCF-v2b5-000042); a link to a family of an earlier clustering shows where its members went.

The families were recomputed on the current corpus (data release R2.0-beta.4): 614,268 listed families, 88.4% of BGCs in one. Family names carry the new tag (GCF-v2b4-000042); a link to a family of either previous clustering shows where its members went.

The families were recomputed on the current corpus (data release R2.0-beta.2): 614,215 listed families, 88.4% of BGCs in one, including the 8,490 GTDB BGCs added on 9 October. Family names now carry the release their clustering first shipped in (GCF-v2b2-000042); a link to a family of the previous clustering shows which current families its members are in.

Taxonomy is now assigned per BGC rather than inherited per sample, lifting coverage from 71% to 99.4% of the 16.5M BGCs (82% resolved to genus or better), with GTDB refreshed to release 232. It flows through the BGC, GCF and sample tables, the metadata filters, GCF taxonomy distributions, and sequence-search results.

Upload a whole GenBank record, not just proteins: hits are grouped per BGC on a paginated table sortable by how many query genes matched, with BLAST-style pairwise alignments, a contig map showing where each BGC's genes align on your query, a map of where the hits come from, and live progress while the search runs.

Matched BGCs can now be viewed on a map with a colour-by dropdown, distribution buckets link straight to the page they name, RiPP annotations are split into typed fields you can filter individually, the GCF filter accepts family names, and NRPS module signatures and Stachelhaus codes support substring matching.

The DIAMOND protein index now covers the full Atlas (16.5M BGCs, 170M proteins) instead of a June snapshot that was missing a quarter of it. Hits are grouped by BGC and link to their Atlas pages with product type, sample, source, GCF and taxonomy, and nucleotide queries are now accepted via blastx.

GOLD ecosystem classification, geographic coordinates, and ENVO biome/feature/material terms now cover ~94% of 860k+ samples, spanning GEM, SPIRE, SMAG, TPMC, OMD, GTDB, MGnify, Microflora Danica, and more.

A new query builder to search all ~16.5M BGCs by their antiSMASH properties — product type, biosynthetic genes, domains, taxonomy, and ecosystem — with live result-distribution charts.

Explore any BGC in an embedded, interactive antiSMASH cluster viewer, with one-click "Open in antiSMASH" links that jump straight to the matching region.

1.76M new BGCs from newly processed MGnify metagenomes — the Atlas now holds ~16.5M BGCs.

A new Genome Collections browser plus 7 MAG catalogs (+541k genomes, +2.1M BGCs), with taxonomy, genome-type, and GOLD-ecosystem breakdowns.

Search the Atlas by protein sequence (DIAMOND) or upload a GenBank file to find related gene cluster families (BiG-SLiCE).

New Analyses pages: biome specificity, ecosystem specificity, and an interactive GCF co-occurrence network.

Added 14K new datasets from environmental sources from the SPIRE database.

BGC Atlas now uses antiSMASH 8. All datasets were updated. New datasets in MGnify were also processed.

Initial release of BGC Atlas with comprehensive metagenomic BGC database, interactive web interface, and advanced search capabilities.