# BGC Atlas data releases

Newest first. Each release is in /downloads/releases/<id>/ (RELEASE.md, manifest.json, SHA256SUMS, bgc_ids_added.txt.gz, bgc_ids_removed.txt.gz). Alphas are not listed.

## R2.0 (2026-10-22)

Against R2.0-beta.6.

| | count | change |
|---|---:|---:|
| BGCs | 16,109,198 |  |
| Complete BGCs | 2,699,040 |  |
| GCFs | 629,872 |  |
| BGCs in a listed GCF | 14,228,788 |  |
| Metagenome samples | 67,067 |  |
| Genomes | 750,901 |  |
| Assemblies | 836,241 |  |

BGC ids: 0 added, 0 removed.

### Changes

The final release of BGC Atlas v2, the data described in the paper. It holds the same BGCs, families, memberships and annotations as R2.0-beta.6; what changes is the family names and the files that only served ids of earlier betas.

- Family names carry the tag `v2`: `GCF-v2-000042`, `GCF-v2-F-400042`. These are the R2.0-beta.6 families `GCF-v2b6-000042` and `GCF-v2b6-F-400042` under their final names. Their uuids (`gcfs.id`) and two-word aliases are unchanged, and a `GCF-v2b6-...` name still opens the same family on the site.
- `db/gcf_legacy` and `db/gcf_successor` are removed. They listed the families of the clusterings replaced during the beta and where their members went; those families' names no longer resolve.
- `bgc_id_redirects.tsv.gz` and `db/bgc_redirect` are removed. They mapped the 356,305 BGC ids withdrawn on 2026-10-05 (MGnify assemblies loaded twice) to the identical BGCs that remain; those ids no longer resolve.
- Unchanged: the BGCs and their files (`all_regions_gbk.tar.gz`, `regions_nuc.fasta`, `regions_prot.faa`, the sequence-search index, `bgc_uuid_source`), and `clustering-v3/` apart from the names in its README.
- From now on, corrections that keep the same BGCs are released as R2.0.1, R2.0.2, ...; added or removed BGCs or a new clustering as R2.1. The pre-releases are retired: their notes, checksums and id lists stay.

### Known issues

- None known.

### Files

- changed: `db/README.txt` (5.3 KB -> 4.7 KB)
- changed: `db/analyses.tsv.gz` (32.1 MB -> 32.1 MB)
- changed: `db/antismash_runs.tsv.gz` (62.6 MB -> 62.6 MB)
- changed: `db/assemblies.tsv.gz` (25.7 MB -> 25.7 MB)
- changed: `db/assembly_analyses.tsv.gz` (36.4 MB -> 36.4 MB)
- changed: `db/assembly_taxonomy.tsv.gz` (43.3 MB -> 43.3 MB)
- changed: `db/bgc_atlas_v2_core.dump` (2.7 GB -> 2.7 GB)
- changed: `db/bgc_atlas_v2_tables_parquet.tar.gz` (3.4 GB -> 3.2 GB)
- changed: `db/bgc_atlas_v2_tables_tsv.tar.gz` (4.1 GB -> 3.9 GB)
- changed: `db/bgc_features.tsv.gz` (1.1 GB -> 1.1 GB)
- changed: `db/bgc_gcf.tsv.gz` (645.3 MB -> 645.3 MB)
- changed: `db/bgc_taxonomy.tsv.gz` (602.5 MB -> 602.5 MB)
- changed: `db/bgcs.tsv.gz` (916.1 MB -> 916.1 MB)
- changed: `db/biomes.tsv.gz` (20.0 KB -> 20.0 KB)
- changed: `db/data_sources.tsv.gz` (1.6 KB -> 1.6 KB)
- changed: `db/gcf_mibig_anchor.tsv.gz` (97.2 KB -> 97.2 KB)
- changed: `db/gcf_study_count.tsv.gz` (14.8 MB -> 14.8 MB)
- changed: `db/gcf_taxonomy_distribution_mv.tsv.gz` (57.7 MB -> 57.7 MB)
- changed: `db/gcfs.tsv.gz` (48.3 MB -> 48.2 MB)
- changed: `db/genomes.tsv.gz` (52.6 MB -> 52.6 MB)
- changed: `db/mgnify_genome_catalogues.tsv.gz` (3.7 KB -> 3.7 KB)
- changed: `db/mgnify_genomes.tsv.gz` (5.0 MB -> 5.0 MB)
- changed: `db/mgnify_publications.tsv.gz` (278.1 KB -> 278.1 KB)
- changed: `db/mgnify_super_studies.tsv.gz` (36.6 KB -> 36.6 KB)
- changed: `db/product_category_map.tsv.gz` (753 B -> 753 B)
- changed: `db/run_analyses.tsv.gz` (1.0 MB -> 1.0 MB)
- changed: `db/run_assemblies.tsv.gz` (2.5 MB -> 2.5 MB)
- changed: `db/run_studies.tsv.gz` (1.6 MB -> 1.6 MB)
- changed: `db/runs.tsv.gz` (3.0 MB -> 3.0 MB)
- changed: `db/sample_analyses.tsv.gz` (36.2 MB -> 36.2 MB)
- changed: `db/sample_assemblies.tsv.gz` (35.3 MB -> 35.3 MB)
- changed: `db/sample_biomes.tsv.gz` (1.8 MB -> 1.8 MB)
- changed: `db/sample_measurements.tsv.gz` (11.2 MB -> 11.2 MB)
- changed: `db/sample_runs.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/samples.tsv.gz` (144.8 MB -> 144.8 MB)
- changed: `db/studies.tsv.gz` (2.9 MB -> 2.9 MB)
- changed: `db/study_bgc_summary.tsv.gz` (4.5 MB -> 4.5 MB)
- changed: `db/study_biomes.tsv.gz` (63 B -> 63 B)
- changed: `db/study_gcf.tsv.gz` (61.7 MB -> 61.7 MB)
- changed: `db/study_samples.tsv.gz` (11.4 MB -> 11.4 MB)
- changed: `db/taxonomies.tsv.gz` (12.4 MB -> 12.4 MB)
- changed: `clustering-v3/README.txt` (1.5 KB -> 1.6 KB)
- removed: `bgc_id_redirects.tsv.gz`
- removed: `db/bgc_redirect.tsv.gz`
- removed: `db/gcf_legacy.tsv.gz`
- removed: `db/gcf_successor.tsv.gz`
- 10 files unchanged

## R2.0-beta.6 (2026-10-14, beta)

Against R2.0-beta.5.

| | count | change |
|---|---:|---:|
| BGCs | 16,109,198 |  |
| Complete BGCs | 2,699,040 |  |
| GCFs | 629,872 | +11,177 |
| BGCs in a listed GCF | 14,228,788 |  |
| Metagenome samples | 67,067 |  |
| Genomes | 750,901 |  |
| Assemblies | 836,241 |  |

BGC ids: 0 added, 0 removed.

### Changes

Gene cluster families corrected and recomputed on R2.0-beta.5's 16,109,198 BGCs, which are unchanged.

- 800,091 families, 629,872 of them listed (373,360 seed, 256,512 fragmented). 14,228,788 BGCs (88.3%) are in a listed family; 2,366,096 are in more than one. 1,328 listed families contain a MIBiG reference. The seed run is the same as R2.0-beta.5's, so the seed families are too, with changed members; the fragmented families are all new.
- Family names carry the tag `v2b6`: `GCF-v2b6-000042`, `GCF-v2b6-F-400042`. The R2.0-beta.5 families (`GCF-v2b5-...`) are replaced without being kept as an earlier clustering: they are not in `db/gcf_legacy`, and their names no longer resolve on the site. `db/gcf_successor` links the v2b4 families (`clustering = 'v2b4-2026-10-12'`) to the v2b6 families directly.
- Per-family tables follow the new families: `db/gcfs`, `db/bgc_gcf`, `db/bgc_features`, `db/gcf_mibig_anchor`, `db/gcf_taxonomy_distribution_mv`. Per-study figures too: % clustered and exclusive GCFs on /studies, `db/study_gcf`, `db/gcf_study_count`, `db/study_bgc_summary`.
- `clustering-v3/`: the two BiG-SLiCE runs of this clustering.
- Unchanged: the BGCs and their files (`all_regions_gbk.tar.gz`, `regions_nuc.fasta`, `regions_prot.faa`, the sequence-search index, `bgc_uuid_source`).

### Known issues

- None known.

### Files

- changed: `db/README.txt` (5.3 KB -> 5.3 KB)
- changed: `db/analyses.tsv.gz` (32.1 MB -> 32.1 MB)
- changed: `db/antismash_runs.tsv.gz` (62.6 MB -> 62.6 MB)
- changed: `db/assemblies.tsv.gz` (25.7 MB -> 25.7 MB)
- changed: `db/assembly_analyses.tsv.gz` (36.4 MB -> 36.4 MB)
- changed: `db/assembly_taxonomy.tsv.gz` (43.3 MB -> 43.3 MB)
- changed: `db/bgc_atlas_v2_core.dump` (2.7 GB -> 2.7 GB)
- changed: `db/bgc_atlas_v2_tables_parquet.tar.gz` (3.4 GB -> 3.4 GB)
- changed: `db/bgc_atlas_v2_tables_tsv.tar.gz` (4.1 GB -> 4.1 GB)
- changed: `db/bgc_features.tsv.gz` (1.1 GB -> 1.1 GB)
- changed: `db/bgc_gcf.tsv.gz` (641.1 MB -> 645.3 MB)
- changed: `db/bgc_redirect.tsv.gz` (14.9 MB -> 14.9 MB)
- changed: `db/bgc_taxonomy.tsv.gz` (602.5 MB -> 602.5 MB)
- changed: `db/bgcs.tsv.gz` (916.1 MB -> 916.1 MB)
- changed: `db/biomes.tsv.gz` (20.0 KB -> 20.0 KB)
- changed: `db/data_sources.tsv.gz` (1.6 KB -> 1.6 KB)
- changed: `db/gcf_legacy.tsv.gz` (81.5 MB -> 81.5 MB)
- changed: `db/gcf_mibig_anchor.tsv.gz` (95.7 KB -> 97.2 KB)
- changed: `db/gcf_study_count.tsv.gz` (14.2 MB -> 14.8 MB)
- changed: `db/gcf_successor.tsv.gz` (110.8 MB -> 120.7 MB)
- changed: `db/gcf_taxonomy_distribution_mv.tsv.gz` (57.1 MB -> 57.7 MB)
- changed: `db/gcfs.tsv.gz` (46.3 MB -> 48.3 MB)
- changed: `db/genomes.tsv.gz` (52.6 MB -> 52.6 MB)
- changed: `db/mgnify_genome_catalogues.tsv.gz` (3.7 KB -> 3.7 KB)
- changed: `db/mgnify_genomes.tsv.gz` (5.0 MB -> 5.0 MB)
- changed: `db/mgnify_publications.tsv.gz` (278.1 KB -> 278.1 KB)
- changed: `db/mgnify_super_studies.tsv.gz` (36.6 KB -> 36.6 KB)
- changed: `db/product_category_map.tsv.gz` (753 B -> 753 B)
- changed: `db/run_analyses.tsv.gz` (1.0 MB -> 1.0 MB)
- changed: `db/run_assemblies.tsv.gz` (2.5 MB -> 2.5 MB)
- changed: `db/run_studies.tsv.gz` (1.6 MB -> 1.6 MB)
- changed: `db/runs.tsv.gz` (3.0 MB -> 3.0 MB)
- changed: `db/sample_analyses.tsv.gz` (36.2 MB -> 36.2 MB)
- changed: `db/sample_assemblies.tsv.gz` (35.3 MB -> 35.3 MB)
- changed: `db/sample_biomes.tsv.gz` (1.8 MB -> 1.8 MB)
- changed: `db/sample_measurements.tsv.gz` (11.2 MB -> 11.2 MB)
- changed: `db/sample_runs.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/samples.tsv.gz` (144.8 MB -> 144.8 MB)
- changed: `db/studies.tsv.gz` (2.9 MB -> 2.9 MB)
- changed: `db/study_bgc_summary.tsv.gz` (4.4 MB -> 4.5 MB)
- changed: `db/study_biomes.tsv.gz` (63 B -> 63 B)
- changed: `db/study_gcf.tsv.gz` (62.4 MB -> 61.7 MB)
- changed: `db/study_samples.tsv.gz` (11.4 MB -> 11.4 MB)
- changed: `db/taxonomies.tsv.gz` (12.4 MB -> 12.4 MB)
- changed: `clustering-v3/README.txt` (1.5 KB -> 1.5 KB)
- changed: `clustering-v3/bigslice-v3-residual.tar.gz` (844.5 MB -> 936.0 MB)
- changed: `clustering-v3/bigslice-v3-seed.tar.gz` (1.2 GB -> 1.2 GB)
- 9 files unchanged

## R2.0-beta.5 (2026-10-13, beta)

Against R2.0-beta.4.

| | count | change |
|---|---:|---:|
| BGCs | 16,109,198 | +124,176 |
| Complete BGCs | 2,699,040 | +28,421 |
| GCFs | 618,695 | +4,427 |
| BGCs in a GCF | 14,232,505 | +103,920 |
| Metagenome samples | 67,067 | +234 |
| Genomes | 750,901 |  |
| Assemblies | 836,241 | +234 |

BGC ids: 124,176 added, 0 removed.

### Changes

124,176 BGCs added and the gene cluster families recomputed on all 16,109,198. No BGC was removed.

- New BGCs: 66,549 from 173 metalog samples (SPIRE assemblies), and two long-read data sources: Great Barrier Reef seawater (`custom_great_barrier_reef`, 38,505 BGCs, PRJEB82623) and the Helgoland 2020 spring bloom (`custom_helgoland_2020`, 19,122 BGCs, PRJEB52999). They are in every sequence file, the sequence-search index and `bgc_uuid_source`.
- Families recomputed with the same method (BiG-SLiCE 2.0 at 0.4, seed and fragmented tiers): 767,255 families, 618,695 listed (373,384 seed, 245,311 fragmented). 14,232,505 BGCs (88.4%) are in a family; 2,429,516 are in more than one. 1,336 listed families contain a MIBiG reference.
- Family names carry the tag `v2b5`: `GCF-v2b5-000042`, `GCF-v2b5-F-400042`. `db/gcf_legacy` adds the v2b4 families (`clustering = 'v2b4-2026-10-12'`); `db/gcf_successor` links them to the v2b5 families.
- `clustering-v3/`: the two BiG-SLiCE runs of this clustering.

### Known issues

- None known.

### Files

- changed: `all_regions_gbk.tar.gz` (166.6 GB -> 167.9 GB)
- changed: `all_regions_gbk.index.tsv.gz` (502.6 MB -> 506.5 MB)
- changed: `regions_nuc.fasta` (164.7 GB -> 166.0 GB)
- changed: `regions_prot.faa` (54.9 GB -> 55.3 GB)
- changed: `bgc_uuid_source.parquet` (521.5 MB -> 527.0 MB)
- changed: `bgc_uuid_source.tsv.gz` (605.0 MB -> 612.4 MB)
- changed: `README_bgc_uuid_source.md` (4.7 KB -> 2.2 KB)
- changed: `db/README.txt` (7.4 KB -> 5.3 KB)
- changed: `db/analyses.tsv.gz` (32.1 MB -> 32.1 MB)
- changed: `db/antismash_runs.tsv.gz` (62.6 MB -> 62.6 MB)
- changed: `db/assemblies.tsv.gz` (25.7 MB -> 25.7 MB)
- changed: `db/assembly_analyses.tsv.gz` (36.4 MB -> 36.4 MB)
- changed: `db/assembly_taxonomy.tsv.gz` (43.3 MB -> 43.3 MB)
- changed: `db/bgc_atlas_v2_core.dump` (2.7 GB -> 2.7 GB)
- changed: `db/bgc_atlas_v2_tables_parquet.tar.gz` (3.3 GB -> 3.4 GB)
- changed: `db/bgc_atlas_v2_tables_tsv.tar.gz` (4.0 GB -> 4.1 GB)
- changed: `db/bgc_features.tsv.gz` (1.0 GB -> 1.1 GB)
- changed: `db/bgc_gcf.tsv.gz` (635.3 MB -> 641.1 MB)
- changed: `db/bgc_redirect.tsv.gz` (14.9 MB -> 14.9 MB)
- changed: `db/bgc_taxonomy.tsv.gz` (598.0 MB -> 602.5 MB)
- changed: `db/bgcs.tsv.gz` (906.0 MB -> 916.1 MB)
- changed: `db/biomes.tsv.gz` (19.5 KB -> 20.0 KB)
- changed: `db/data_sources.tsv.gz` (1.4 KB -> 1.6 KB)
- changed: `db/gcf_legacy.tsv.gz` (54.3 MB -> 81.5 MB)
- changed: `db/gcf_mibig_anchor.tsv.gz` (96.0 KB -> 95.7 KB)
- changed: `db/gcf_study_count.tsv.gz` (14.1 MB -> 14.2 MB)
- changed: `db/gcf_successor.tsv.gz` (73.6 MB -> 110.8 MB)
- changed: `db/gcf_taxonomy_distribution_mv.tsv.gz` (56.7 MB -> 57.1 MB)
- changed: `db/gcfs.tsv.gz` (46.0 MB -> 46.3 MB)
- changed: `db/genomes.tsv.gz` (52.6 MB -> 52.6 MB)
- changed: `db/mgnify_genome_catalogues.tsv.gz` (3.7 KB -> 3.7 KB)
- changed: `db/mgnify_genomes.tsv.gz` (5.0 MB -> 5.0 MB)
- changed: `db/mgnify_publications.tsv.gz` (278.1 KB -> 278.1 KB)
- changed: `db/mgnify_super_studies.tsv.gz` (36.6 KB -> 36.6 KB)
- changed: `db/product_category_map.tsv.gz` (753 B -> 753 B)
- changed: `db/run_analyses.tsv.gz` (1.0 MB -> 1.0 MB)
- changed: `db/run_assemblies.tsv.gz` (2.5 MB -> 2.5 MB)
- changed: `db/run_studies.tsv.gz` (1.6 MB -> 1.6 MB)
- changed: `db/runs.tsv.gz` (3.0 MB -> 3.0 MB)
- changed: `db/sample_analyses.tsv.gz` (36.2 MB -> 36.2 MB)
- changed: `db/sample_assemblies.tsv.gz` (35.3 MB -> 35.3 MB)
- changed: `db/sample_biomes.tsv.gz` (1.8 MB -> 1.8 MB)
- changed: `db/sample_measurements.tsv.gz` (11.1 MB -> 11.2 MB)
- changed: `db/sample_runs.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/samples.tsv.gz` (144.8 MB -> 144.8 MB)
- changed: `db/studies.tsv.gz` (2.8 MB -> 2.9 MB)
- changed: `db/study_bgc_summary.tsv.gz` (4.4 MB -> 4.4 MB)
- changed: `db/study_biomes.tsv.gz` (63 B -> 63 B)
- changed: `db/study_gcf.tsv.gz` (61.9 MB -> 62.4 MB)
- changed: `db/study_samples.tsv.gz` (11.4 MB -> 11.4 MB)
- changed: `db/taxonomies.tsv.gz` (12.4 MB -> 12.4 MB)
- changed: `clustering-v3/README.txt` (1.8 KB -> 1.5 KB)
- changed: `clustering-v3/bigslice-v3-residual.tar.gz` (840.8 MB -> 844.5 MB)
- changed: `clustering-v3/bigslice-v3-seed.tar.gz` (1.2 GB -> 1.2 GB)
- 2 files unchanged

## R2.0-beta.4 (2026-10-12, beta)

Against R2.0-beta.3.

| | count | change |
|---|---:|---:|
| BGCs | 15,985,022 |  |
| Complete BGCs | 2,670,619 |  |
| GCFs | 614,268 | +53 |
| BGCs in a GCF | 14,128,585 | +384,475 |
| Metagenome samples | 66,833 |  |
| Genomes | 750,901 |  |
| Assemblies | 836,007 |  |

BGC ids: 0 added, 0 removed.

### Changes

Gene cluster families recomputed on R2.0-beta.3's BGCs, which are unchanged. This gives a family again to the 387,199 MGnify BGCs that R2.0-beta.3 re-analysed and left without one.

- The families were recomputed with the same method (BiG-SLiCE 2.0 at 0.4, seed and fragmented tiers): 762,045 families, 614,268 of them listed (370,690 seed, 243,578 fragmented). 14,128,585 BGCs (88.4%) are in a family; 2,381,240 are in more than one. 1,346 listed families contain a MIBiG reference.
- Family names carry the new tag: `GCF-v2b4-000042`, `GCF-v2b4-F-400042`. Families are numbered afresh in each clustering, so `GCF-v2b4-000042` is unrelated to `GCF-v2b2-000042`. `db/gcf_legacy` now lists the families of both previous clusterings (column `clustering`: `v3-2026-08`, `v2b2-2026-10-09`), and `db/gcf_successor` says which families of the next clustering each one's members are in: the v2b2 families point at the v2b4 families, the August 2026 families still at the v2b2 families. On the site, a link to a previous family shows the same, one clustering at a time.
- Per-family tables follow the new families: `db/gcfs`, `db/bgc_gcf`, `db/bgc_features`, `db/gcf_mibig_anchor`, `db/gcf_taxonomy_distribution_mv`. `db/bgc_features` again has a row for every BGC.
- Per-study figures follow the new families: % clustered and exclusive GCFs on /studies, `db/study_gcf`, `db/gcf_study_count`, `db/study_bgc_summary`.
- `clustering-v3/`: the two BiG-SLiCE runs of this clustering, replacing R2.0-beta.2's.
- Unchanged: the BGCs and their files (`all_regions_gbk.tar.gz`, `regions_nuc.fasta`, `regions_prot.faa`, the sequence-search index, `bgc_uuid_source`).

### Known issues

- None known.

### Files

- changed: `db/README.txt` (7.1 KB -> 7.4 KB)
- changed: `db/analyses.tsv.gz` (32.1 MB -> 32.1 MB)
- changed: `db/antismash_runs.tsv.gz` (62.6 MB -> 62.6 MB)
- changed: `db/assemblies.tsv.gz` (25.7 MB -> 25.7 MB)
- changed: `db/assembly_analyses.tsv.gz` (36.4 MB -> 36.4 MB)
- changed: `db/assembly_taxonomy.tsv.gz` (43.3 MB -> 43.3 MB)
- changed: `db/bgc_atlas_v2_core.dump` (2.7 GB -> 2.7 GB)
- changed: `db/bgc_atlas_v2_tables_parquet.tar.gz` (3.2 GB -> 3.3 GB)
- changed: `db/bgc_atlas_v2_tables_tsv.tar.gz` (3.9 GB -> 4.0 GB)
- changed: `db/bgc_features.tsv.gz` (1.0 GB -> 1.0 GB)
- changed: `db/bgc_gcf.tsv.gz` (617.3 MB -> 635.3 MB)
- changed: `db/bgc_redirect.tsv.gz` (14.9 MB -> 14.9 MB)
- changed: `db/bgc_taxonomy.tsv.gz` (598.0 MB -> 598.0 MB)
- changed: `db/bgcs.tsv.gz` (906.0 MB -> 906.0 MB)
- changed: `db/biomes.tsv.gz` (19.5 KB -> 19.5 KB)
- changed: `db/data_sources.tsv.gz` (1.4 KB -> 1.4 KB)
- changed: `db/gcf_legacy.tsv.gz` (27.0 MB -> 54.3 MB)
- changed: `db/gcf_mibig_anchor.tsv.gz` (94.7 KB -> 96.0 KB)
- changed: `db/gcf_study_count.tsv.gz` (13.9 MB -> 14.1 MB)
- changed: `db/gcf_successor.tsv.gz` (37.0 MB -> 73.6 MB)
- changed: `db/gcf_taxonomy_distribution_mv.tsv.gz` (56.1 MB -> 56.7 MB)
- changed: `db/gcfs.tsv.gz` (45.9 MB -> 46.0 MB)
- changed: `db/genomes.tsv.gz` (52.6 MB -> 52.6 MB)
- changed: `db/mgnify_genome_catalogues.tsv.gz` (3.7 KB -> 3.7 KB)
- changed: `db/mgnify_genomes.tsv.gz` (5.0 MB -> 5.0 MB)
- changed: `db/mgnify_publications.tsv.gz` (278.1 KB -> 278.1 KB)
- changed: `db/mgnify_super_studies.tsv.gz` (36.6 KB -> 36.6 KB)
- changed: `db/product_category_map.tsv.gz` (753 B -> 753 B)
- changed: `db/run_analyses.tsv.gz` (1.0 MB -> 1.0 MB)
- changed: `db/run_assemblies.tsv.gz` (2.5 MB -> 2.5 MB)
- changed: `db/run_studies.tsv.gz` (1.6 MB -> 1.6 MB)
- changed: `db/runs.tsv.gz` (3.0 MB -> 3.0 MB)
- changed: `db/sample_analyses.tsv.gz` (36.2 MB -> 36.2 MB)
- changed: `db/sample_assemblies.tsv.gz` (35.3 MB -> 35.3 MB)
- changed: `db/sample_biomes.tsv.gz` (1.8 MB -> 1.8 MB)
- changed: `db/sample_measurements.tsv.gz` (11.1 MB -> 11.1 MB)
- changed: `db/sample_runs.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/samples.tsv.gz` (144.8 MB -> 144.8 MB)
- changed: `db/studies.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/study_bgc_summary.tsv.gz` (4.4 MB -> 4.4 MB)
- changed: `db/study_biomes.tsv.gz` (63 B -> 63 B)
- changed: `db/study_gcf.tsv.gz` (61.1 MB -> 61.9 MB)
- changed: `db/study_samples.tsv.gz` (11.4 MB -> 11.4 MB)
- changed: `db/taxonomies.tsv.gz` (12.4 MB -> 12.4 MB)
- changed: `clustering-v3/README.txt` (1.8 KB -> 1.8 KB)
- changed: `clustering-v3/bigslice-v3-residual.tar.gz` (840.5 MB -> 840.8 MB)
- changed: `clustering-v3/bigslice-v3-seed.tar.gz` (1.2 GB -> 1.2 GB)
- 9 files unchanged

## R2.0-beta.3 (2026-10-12, beta)

Against R2.0-beta.2.

| | count | change |
|---|---:|---:|
| BGCs | 15,985,022 | +7,638 |
| Complete BGCs | 2,670,619 | +618 |
| GCFs | 614,215 |  |
| BGCs in a GCF | 13,744,110 | -379,522 |
| Metagenome samples | 66,833 |  |
| Genomes | 750,901 |  |
| Assemblies | 836,007 |  |

BGC ids: 10,610 added, 2,972 removed.

### Changes

The 20,189 MGnify assemblies analysed in 2026 were re-analysed. Their genes had been called with prodigal in single-genome mode, which is meant for one genome, not a metagenome. Contigs of at least 20 kb were re-analysed with prodigal in metagenome mode (same metaSMASH 1.0.3, antiSMASH 8.0.4 core); regions on shorter contigs, already gene-called in metagenome mode, were kept from the original run. 6,307 assemblies gained or lost a region; in the rest, regions on long contigs may have new boundaries or genes.

- 15,985,022 BGCs (+7,638): 10,610 regions are new and 2,972 are gone (`bgc_ids_added.txt.gz`, `bgc_ids_removed.txt.gz`). Removed ids are not redirected.
- A BGC id is derived from the region's name, so a region that kept its name kept its id: 1,760,608 did. 1,384,019 of them are unchanged; 376,589 have new coordinates or a different set of proteins.
- The 387,199 new and changed BGCs are not in a gene cluster family until the families are next recomputed: their earlier placement described different genes. 13,744,110 BGCs (86.0%) are in a family, down from 88.4%; 2,313,679 are in more than one. 3,307 listed families lost all their members; they stay in `db/gcfs` with `n_bgc` 0 and are hidden on the site.
- Taxonomy of the new and changed BGCs recomputed as before (mmseqs2 on the region's proteins against GTDB R232): 387,003 classified; the 196 that could not be assigned a domain have none. Of the changed BGCs, 99.8% kept the same lineage.
- Changed files: `all_regions_gbk.tar.gz` with its index and `extract_gbk.py`, `regions_nuc.fasta`, `regions_prot.faa` and the sequence-search index built from it, `bgc_uuid_source` (and its README), and `db/`. Per-study figures follow (`db/study_gcf`, `db/gcf_study_count`, `db/study_bgc_summary`).
- `clustering-v3/` is unchanged: it describes the R2.0-beta.2 families.

### Known issues

- 387,199 MGnify BGCs have no family (`db/bgc_gcf`) and no row in the clustering's per-BGC table (`db/bgc_features`) until the families are recomputed.

### Files

- changed: `all_regions_gbk.tar.gz` (166.5 GB -> 166.6 GB)
- changed: `all_regions_gbk.index.tsv.gz` (502.4 MB -> 502.6 MB)
- changed: `regions_nuc.fasta` (164.6 GB -> 164.7 GB)
- changed: `regions_prot.faa` (54.8 GB -> 54.9 GB)
- changed: `bgc_uuid_source.parquet` (522.8 MB -> 521.5 MB)
- changed: `bgc_uuid_source.tsv.gz` (604.6 MB -> 605.0 MB)
- changed: `README_bgc_uuid_source.md` (4.3 KB -> 4.7 KB)
- changed: `db/README.txt` (7.1 KB -> 7.1 KB)
- changed: `db/analyses.tsv.gz` (32.1 MB -> 32.1 MB)
- changed: `db/antismash_runs.tsv.gz` (62.6 MB -> 62.6 MB)
- changed: `db/assemblies.tsv.gz` (25.7 MB -> 25.7 MB)
- changed: `db/assembly_analyses.tsv.gz` (36.4 MB -> 36.4 MB)
- changed: `db/assembly_taxonomy.tsv.gz` (43.3 MB -> 43.3 MB)
- changed: `db/bgc_atlas_v2_core.dump` (2.7 GB -> 2.7 GB)
- changed: `db/bgc_atlas_v2_tables_parquet.tar.gz` (3.2 GB -> 3.2 GB)
- changed: `db/bgc_atlas_v2_tables_tsv.tar.gz` (3.9 GB -> 3.9 GB)
- changed: `db/bgc_features.tsv.gz` (1.0 GB -> 1.0 GB)
- changed: `db/bgc_gcf.tsv.gz` (636.7 MB -> 617.3 MB)
- changed: `db/bgc_redirect.tsv.gz` (14.9 MB -> 14.9 MB)
- changed: `db/bgc_taxonomy.tsv.gz` (598.2 MB -> 598.0 MB)
- changed: `db/bgcs.tsv.gz` (904.1 MB -> 906.0 MB)
- changed: `db/biomes.tsv.gz` (19.5 KB -> 19.5 KB)
- changed: `db/data_sources.tsv.gz` (1.4 KB -> 1.4 KB)
- changed: `db/gcf_legacy.tsv.gz` (27.0 MB -> 27.0 MB)
- changed: `db/gcf_mibig_anchor.tsv.gz` (94.7 KB -> 94.7 KB)
- changed: `db/gcf_study_count.tsv.gz` (14.1 MB -> 13.9 MB)
- changed: `db/gcf_successor.tsv.gz` (37.0 MB -> 37.0 MB)
- changed: `db/gcf_taxonomy_distribution_mv.tsv.gz` (56.7 MB -> 56.1 MB)
- changed: `db/gcfs.tsv.gz` (46.0 MB -> 45.9 MB)
- changed: `db/genomes.tsv.gz` (52.6 MB -> 52.6 MB)
- changed: `db/mgnify_genome_catalogues.tsv.gz` (3.7 KB -> 3.7 KB)
- changed: `db/mgnify_genomes.tsv.gz` (5.0 MB -> 5.0 MB)
- changed: `db/mgnify_publications.tsv.gz` (278.1 KB -> 278.1 KB)
- changed: `db/mgnify_super_studies.tsv.gz` (36.6 KB -> 36.6 KB)
- changed: `db/product_category_map.tsv.gz` (753 B -> 753 B)
- changed: `db/run_analyses.tsv.gz` (1.0 MB -> 1.0 MB)
- changed: `db/run_assemblies.tsv.gz` (2.5 MB -> 2.5 MB)
- changed: `db/run_studies.tsv.gz` (1.6 MB -> 1.6 MB)
- changed: `db/runs.tsv.gz` (3.0 MB -> 3.0 MB)
- changed: `db/sample_analyses.tsv.gz` (36.2 MB -> 36.2 MB)
- changed: `db/sample_assemblies.tsv.gz` (35.3 MB -> 35.3 MB)
- changed: `db/sample_biomes.tsv.gz` (1.8 MB -> 1.8 MB)
- changed: `db/sample_measurements.tsv.gz` (11.1 MB -> 11.1 MB)
- changed: `db/sample_runs.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/samples.tsv.gz` (144.8 MB -> 144.8 MB)
- changed: `db/studies.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/study_bgc_summary.tsv.gz` (4.4 MB -> 4.4 MB)
- changed: `db/study_biomes.tsv.gz` (63 B -> 63 B)
- changed: `db/study_gcf.tsv.gz` (61.6 MB -> 61.1 MB)
- changed: `db/study_samples.tsv.gz` (11.4 MB -> 11.4 MB)
- changed: `db/taxonomies.tsv.gz` (12.4 MB -> 12.4 MB)
- 5 files unchanged

## R2.0-beta.2 (2026-10-09, beta)

Against R2.0-beta.1.

| | count | change |
|---|---:|---:|
| BGCs | 15,977,384 |  |
| Complete BGCs | 2,670,001 |  |
| GCFs | 614,215 | -2,535 |
| BGCs in a GCF | 14,123,632 | +1,101 |
| Metagenome samples | 66,833 |  |
| Genomes | 750,901 |  |
| Assemblies | 836,007 |  |

BGC ids: 0 added, 0 removed.

### Changes

Gene cluster families recomputed. The BGCs are those of R2.0-beta.1, unchanged; what changed is how they are grouped, plus the database tables listed below.

- The families were recomputed with the same method (BiG-SLiCE 2.0 at 0.4, seed and fragmented tiers) on this release's BGCs, including the 8,490 GTDB BGCs that R2.0-beta.1 added without a family: 762,124 families, 614,215 of them listed (370,650 seed, 243,565 fragmented). 14,123,632 BGCs (88.4%) are in a family; 2,402,552 are in more than one. 1,335 listed families contain a MIBiG reference.
- Family names now carry the release their clustering first shipped in: `GCF-v2b2-000042`, `GCF-v2b2-F-400042`. Families are numbered afresh in each clustering, so `GCF-v2b2-000042` is unrelated to R2.0-beta.1's `GCF-000042`. The previous families are listed in `db/gcf_legacy`, and `db/gcf_successor` says which new families each one's members are in (see FAMILY NAMES AND THE PREVIOUS CLUSTERING in `db/README.txt`). On the site, a link to a previous family shows the same.
- `db/gcf_hierarchy` (family super-groups) is gone: the new clustering has no such step.
- `clustering-v3/`: the two new BiG-SLiCE runs. They place new BGCs into these families like the previous runs did, but their per-gene tables are empty, so BiG-SLiCE's report viewer and re-running at another threshold no longer work from them (`clustering-v3/README.txt`).
- Per-study figures follow the new families: % clustered and exclusive GCFs on /studies, `db/study_gcf`, `db/gcf_study_count`, `db/study_bgc_summary`.
- New table `db/genomes.tsv.gz` (also in the TSV and Parquet bundles): one row per genome (isolate, MAG, SAG) with CheckM2 completeness and contamination, the MIMAG label, the source of each genome's values (`checkm_source`: our CheckM2 1.1.0 run, or GTDB R232's published CheckM2), lineage and BGC count. `db/README.txt` documents it under GENOME QUALITY.

### Known issues

- None known.

### Files

- changed: `db/README.txt` (4.5 KB -> 7.1 KB)
- changed: `db/analyses.tsv.gz` (32.1 MB -> 32.1 MB)
- changed: `db/antismash_runs.tsv.gz` (62.6 MB -> 62.6 MB)
- changed: `db/assemblies.tsv.gz` (25.7 MB -> 25.7 MB)
- changed: `db/assembly_analyses.tsv.gz` (36.4 MB -> 36.4 MB)
- changed: `db/assembly_taxonomy.tsv.gz` (43.3 MB -> 43.3 MB)
- changed: `db/bgc_atlas_v2_core.dump` (2.7 GB -> 2.7 GB)
- changed: `db/bgc_atlas_v2_tables_parquet.tar.gz` (3.1 GB -> 3.2 GB)
- changed: `db/bgc_atlas_v2_tables_tsv.tar.gz` (3.8 GB -> 3.9 GB)
- changed: `db/bgc_features.tsv.gz` (1.0 GB -> 1.0 GB)
- changed: `db/bgc_gcf.tsv.gz` (636.7 MB -> 636.7 MB)
- changed: `db/bgc_redirect.tsv.gz` (14.9 MB -> 14.9 MB)
- changed: `db/bgc_taxonomy.tsv.gz` (598.2 MB -> 598.2 MB)
- changed: `db/bgcs.tsv.gz` (904.1 MB -> 904.1 MB)
- changed: `db/biomes.tsv.gz` (19.5 KB -> 19.5 KB)
- changed: `db/data_sources.tsv.gz` (1.4 KB -> 1.4 KB)
- changed: `db/gcf_mibig_anchor.tsv.gz` (96.1 KB -> 94.7 KB)
- changed: `db/gcf_study_count.tsv.gz` (14.1 MB -> 14.1 MB)
- changed: `db/gcf_taxonomy_distribution_mv.tsv.gz` (56.5 MB -> 56.7 MB)
- changed: `db/gcfs.tsv.gz` (46.1 MB -> 46.0 MB)
- changed: `db/mgnify_genome_catalogues.tsv.gz` (3.7 KB -> 3.7 KB)
- changed: `db/mgnify_genomes.tsv.gz` (5.0 MB -> 5.0 MB)
- changed: `db/mgnify_publications.tsv.gz` (278.1 KB -> 278.1 KB)
- changed: `db/mgnify_super_studies.tsv.gz` (36.6 KB -> 36.6 KB)
- changed: `db/product_category_map.tsv.gz` (753 B -> 753 B)
- changed: `db/run_analyses.tsv.gz` (1.0 MB -> 1.0 MB)
- changed: `db/run_assemblies.tsv.gz` (2.5 MB -> 2.5 MB)
- changed: `db/run_studies.tsv.gz` (1.6 MB -> 1.6 MB)
- changed: `db/runs.tsv.gz` (3.0 MB -> 3.0 MB)
- changed: `db/sample_analyses.tsv.gz` (36.2 MB -> 36.2 MB)
- changed: `db/sample_assemblies.tsv.gz` (35.3 MB -> 35.3 MB)
- changed: `db/sample_biomes.tsv.gz` (1.8 MB -> 1.8 MB)
- changed: `db/sample_measurements.tsv.gz` (11.1 MB -> 11.1 MB)
- changed: `db/sample_runs.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/samples.tsv.gz` (144.8 MB -> 144.8 MB)
- changed: `db/studies.tsv.gz` (2.8 MB -> 2.8 MB)
- changed: `db/study_bgc_summary.tsv.gz` (4.2 MB -> 4.4 MB)
- changed: `db/study_biomes.tsv.gz` (63 B -> 63 B)
- changed: `db/study_gcf.tsv.gz` (61.7 MB -> 61.6 MB)
- changed: `db/study_samples.tsv.gz` (11.4 MB -> 11.4 MB)
- changed: `db/taxonomies.tsv.gz` (12.4 MB -> 12.4 MB)
- changed: `clustering-v3/README.txt` (1.3 KB -> 1.8 KB)
- changed: `clustering-v3/bigslice-v3-residual.tar.gz` (6.1 GB -> 840.5 MB)
- changed: `clustering-v3/bigslice-v3-seed.tar.gz` (13.6 GB -> 1.2 GB)
- added: `db/gcf_legacy.tsv.gz` (27.0 MB)
- added: `db/gcf_successor.tsv.gz` (37.0 MB)
- added: `db/genomes.tsv.gz` (52.6 MB)
- removed: `db/gcf_hierarchy.tsv.gz`
- 9 files unchanged

## R2.0-beta.1 (2026-10-09, beta)

The first release.

### Changes

The first versioned release. It is the corpus as of 2026-10-09:

- 46,098 genome-derived assemblies removed (166,457 BGCs): duplicate copies of the same genome, and genomes below MIMAG medium quality. Their BGC ids now return 404.
- 1,204 GTDB genomes whose antiSMASH analysis had failed were re-analysed: 8,490 BGCs added.
- Genome quality is CheckM2 for every genome (750,899 of 750,901 genome-collection samples), with its provenance per genome: `samples.tsv.gz` has `checkm_completeness`, `checkm_contamination`, `genome_quality` (MIMAG high or medium), `checkm_tool` and `checkm_source`, which says whether the values are from our CheckM2 1.1.0 run (500,195 genomes) or GTDB R232's published CheckM2 (250,704).
- `regions_nuc.fasta`, `regions_prot.faa` and `all_regions_gbk.tar.gz` now hold every BGC in the release. Before, 86 regions were missing from all three and 6,416 more from both FASTAs.

### Known issues

- Gene cluster families have not yet been recomputed after the removals above. Until the next release, some GCF member counts and representatives in `db/` describe the corpus before the removals, and the 8,490 added BGCs belong to no GCF.