Reference database snapshot required by funannotate v1.8.17

funannotate database snapshot of 8 October 2026, for funannotate 1.8.17

This archive holds a complete funannotate database folder, as written by
"funannotate setup -i all -b dikarya --wget" on 8 October 2026 with funannotate 1.8.17 (bioconda
package funannotate-1.8.17-pyhdfd78af_5; Python 3.8.20, DIAMOND 2.2.4, HMMER 3.4, AUGUSTUS 3.5.0).
It is made for version 2 of the FunFlux workflow (in preparation;
https://github.com/iLivius/FunFlux), which downloads this archive, checks it against its checksum
and unpacks it for funannotate predict and funannotate annotate. FunFlux version 1 uses the
previous version of this record. The archive contains one folder, funannotate_db_2026.10.08/. The file funannotate_db_1.8.17_2026-10-08.sha256
holds its SHA-256 checksum; funannotate_db_1.8.17_2026-10-08.versions.tsv lists the contents below
as a table.

Contents (version, date and record count as funannotate recorded them; source = download address)
- UniProtKB/Swiss-Prot: release 2026_03 of 2026-09-02, 575,748 sequences. Source: ftp.ebi.ac.uk, UniProt current release.
- Pfam-A: release 38.2 of 2026-01, 30,134 families. Source: ftp.ebi.ac.uk, Pfam current release.
- InterPro (interpro.xml and the entry list): release 110.0 of 2026-09-03, 54,922 entries. Source: ftp.ebi.ac.uk, InterPro current release.
- Gene Ontology (go.obo): release 2026-07-26, 48,340 terms. Source: purl.obolibrary.org/obo/go.obo (served by current.geneontology.org).
- MEROPS (meropsscan.lib): 5,098 sequences. Source: ftp.ebi.ac.uk, MEROPS current release. funannotate 1.8.17 gives it the fixed label "12.5, 2023-01-19" (written in its code, not read from the file); the file is identical to the one in the previous version of this record.
- dbCAN: 699 CAZyme HMMs from dbCAN-HMMdb-V11.txt, with CAZyDB.08062022.fam-activities.txt and the readme. Source: pro.unl.edu/dbCAN2 (Databases/V11). funannotate labels it "14.0, 2025-08-26" from the first lines of that readme, which describe HMMdb release 14.0 with 875 HMMs; the HMM file it downloads holds 699.
- MIBiG protein sequences (mibig_prot_seqs_1.4.fasta): version 1.4, 31,023 sequences. Source: dl.secondarymetabolites.org.
- gene2product (curated gene names and product descriptions): version 1.99 of 2026-05-30, 34,505 records (funannotate's count, which includes the file's column-header line). Source: github.com/nextgenusfs/gene2product.
- Repeat-protein library: version 1.0, 11,950 sequences. Source: funannotate GitHub release "db-mirror".
- BUSCO outgroups: version 1.0, BUSCO sequences of 6 species (the index counts 8 files). Source: funannotate GitHub release "db-mirror".
- BUSCO dikarya_odb9: created 2016-02-13, 1,312 BUSCOs. Source: funannotate GitHub release "db-mirror".
- Trained species: AUGUSTUS parameters of 166 species, copied from AUGUSTUS 3.5.0 (bioconda).
For the repeat library, MIBiG and the BUSCO outgroups funannotate records the day of the build
(2026-10-08) as the date.

Changes from the previous version (funannotate_db_1.8.15_2025-12-15)
- Newer releases: UniProtKB/Swiss-Prot 2025_04 -> 2026_03, Pfam 38.0 -> 38.2, InterPro 107.0 -> 110.0,
  Gene Ontology 2025-10-10 -> 2026-07-26, gene2product 1.97 -> 1.99.
- Same content as before: dbCAN, MEROPS, MIBiG, the repeat-protein library, the BUSCO outgroups,
  dikarya_odb9 and the 166 trained species. The MEROPS label changed from 12.0 to 12.5 and its file
  from merops_scan.lib to meropsscan.lib because funannotate 1.8.17 names them so; the file content
  is identical. The DIAMOND files of the unchanged databases differ from the previous ones only in
  the DIAMOND build number written in their header.
- Built with funannotate 1.8.17 (the previous version's name gives 1.8.15) and DIAMOND 2.2.4.

How it was built
In FunFlux's funannotate 1.8.17 conda environment, with a build script written for FunFlux
version 2 (in preparation). It runs the setup command above, checks that every database is listed
in funannotate-db-info.txt with an existing file and more than 0 records and that the dikarya set
and the trained species are present, writes the version table, packs the folder and writes the
checksum. Before packing, the absolute paths of the build computer that funannotate stores in
funannotate-db-info.txt and trained_species/*/info.json were replaced with the placeholder
/path/to/funannotate_db, and the archive records owner and group 0.
--wget makes funannotate download with wget: on the day of the build geneontology.org refused
funannotate's own Python downloader (HTTP 403). funannotate setup reads its download addresses from
funannotate's GitHub repository at run time.

How to use it without FunFlux
mkdir -p /path/to/funannotate_db
tar -xzf funannotate_db_1.8.17_2026-10-08.tar.gz --strip-components=1 -C /path/to/funannotate_db
cd /path/to/funannotate_db
sed -i "s#`head -n 1 funannotate-db-info.txt | cut -f 3 | xargs dirname`#`pwd`#g" funannotate-db-info.txt trained_species/*/info.json
export FUNANNOTATE_DB=/path/to/funannotate_db
funannotate database

The database index (funannotate-db-info.txt) and the info.json file of every trained species store
absolute paths, here the placeholder /path/to/funannotate_db, and funannotate predict reads a
species' AUGUSTUS parameters from the path in its info.json. The sed line rewrites those paths to
the new place; the new folder's path must not contain & or #, which sed would read as
instructions. FunFlux version 2 (in preparation) rewrites these paths itself and checks that every
rewritten path exists. The last command should list ten databases.
Use funannotate 1.8.17; the DIAMOND databases are in DIAMOND's database format 3.

Licences
This archive repackages files downloaded from the sources above; each component keeps the licence
and the citation requirements of its source. As recorded in the licensing notes of FunFlux
version 2 (in preparation) on 8 October 2026:
- UniProtKB/Swiss-Prot: CC BY 4.0; cite UniProt (https://ftp.uniprot.org/pub/databases/uniprot/LICENSE).
- Pfam and InterPro: CC0 1.0 (https://interpro-documentation.readthedocs.io/en/latest/license.html).
- Gene Ontology: CC BY 4.0; cite the Gene Ontology (https://geneontology.org/docs/go-citation-policy/).
- MEROPS: GNU Library General Public License; share-alike (https://www.ebi.ac.uk/merops/about/availability.shtml).
- dbCAN: none stated. The site states no terms for the downloads; its only copyright notice is
  "Copyright 2012-now © YIN LAB, UNL. All rights reserved." (https://pro.unl.edu/dbCAN2).
- MIBiG: CC BY 4.0 according to the Bioregistry entry (https://bioregistry.io/mibig); not confirmed
  by MIBiG: its data repository on GitHub shows no licence, and its website could not be read.
- BUSCO dikarya_odb9: CC BY-ND 4.0, from the BUSCO site's licence statement, which names no dataset
  version (https://busco.ezlab.org/); modified datasets may not be shared.
- BUSCO outgroups, trained AUGUSTUS species, gene2product, repeat-protein library: not checked.
funannotate itself (BSD-2-Clause, https://github.com/nextgenusfs/funannotate) wrote the folder; the
archive contains no funannotate program code.