Reference Database Notices
Version 2026-10 — Effective October 8, 2026
This page is an appendix to our Open-Source Software Notices. It enumerates the third-party reference databases that Ploid hosts and makes readable by workflow runs — the detail behind Section 12 of that document.
Why these are listed separately
Most public biological data reaches an analysis the way Section 12 describes: the analysis queries the publisher directly, the content lands in the customer's own sandbox, and Ploid keeps no copy.
The databases below work differently. A taxonomic classifier or a completeness estimator cannot run without a reference database, and many of these are tens to hundreds of gigabytes — far too large to retrieve on every run. So Ploid keeps a copy in its own storage and makes it readable by the compute that needs it. That copy is ours. Hosting it is a form of redistribution, and this page is the disclosure that goes with it.
Each database listed here is the property of its respective copyright holders and is used under the terms of its own license. Nothing in this document grants you any right in these databases beyond what those licenses provide.
What is hosted, and under what terms
Ploid hosts a database only where the terms permit that use. No database whose terms restrict commercial use is hosted. Availability varies by deployment: not every installation of Ploid offers every database below, and some offer none.
| Database | Contents | License | What it requires |
|---|---|---|---|
| Genome Taxonomy Database (GTDB) reference package, release 226, for GTDB-Tk | Reference genomes and taxonomy used to classify metagenome-assembled genomes | CC BY-SA 4.0 | Attribution and share-alike: taxonomic datasets you derive from it and then distribute must be licensed on the same terms. |
| sourmash sketch databases, GTDB release 214 | Sketch signatures and lineage table built from GTDB | CC BY-SA 4.0, inherited from GTDB | The same attribution and share-alike obligations as GTDB itself. |
| sourmash sketch databases, GTDB release 226 | Sketch signatures and lineage table built from GTDB | CC BY-SA 4.0, inherited from GTDB | The same attribution and share-alike obligations as GTDB itself. |
| Kraken2 PlusPF index (2026-02-26) | k-mer index over NCBI RefSeq archaea, bacteria, viral, plasmid, human and UniVec_Core, plus protozoa and fungi | No license asserted by the index distributor; source data carries no NCBI restriction | Attribution to NCBI is customary. See the caveat below on individual records. |
| Kaiju index over the RefSeq non-redundant protein set (2024-08-13) | Protein-level index with the NCBI taxonomy it requires | No license asserted by the index distributor; source data carries no NCBI restriction | Attribution to NCBI is customary. See the caveat below on individual records. |
| eggNOG orthology database (5.0.2), for eggNOG-mapper | Orthologous groups, Gene Ontology terms, PFAM domains, protein names, BiGG reactions and COG functional categories | CC BY 3.0 | Attribution. |
| eggNOG sequence databases (5.0.2), scoped by taxonomy | Protein sequences drawn from the eggNOG set above and restricted to one clade — bacteria, archaea, eukaryotes, fungi, animals, or green plants — searched alongside the orthology database rather than instead of it | CC BY 3.0, inherited from eggNOG | Attribution. |
| BUSCO lineage datasets (odb12), for BUSCO | Sets of near-universal single-copy orthologs, one per lineage across eukaryotes, bacteria and archaea, used to measure how complete a genome assembly or annotation is | CC BY-ND 4.0 | Attribution, and the publisher requires that any publication or product using the datasets cite its paper (below). Hosted exactly as published, as the no-derivatives term requires: modified versions may not be distributed. |
| CheckM reference data (2015-01-16), for CheckM | Marker gene sets and a reference genome tree, used to estimate the completeness and contamination of genomes | CC BY 4.0; its TIGRFAMs models are CC BY-SA 4.0 | Attribution, and share-alike for the TIGRFAMs models if you distribute adapted versions of them. |
| Pfam-A profile HMMs (38.2) | Protein family and domain models | CC0 1.0 | None: the database is dedicated to the public domain. |
| Rfam covariance models (15.1) | RNA family models | CC0 1.0 | None. Rfam asks that work using it cite the database. |
| Dfam curated transposable element families (4.0) | Repeat family models and consensus sequences, used to annotate and mask repeats | CC0 1.0 | None: the database is dedicated to the public domain. |
| Greengenes2 training sets for DADA2 (2024.09) | 16S rRNA reference sequences with taxonomy, for bacteria and archaea | CC BY 4.0; built from Greengenes2, whose taxonomy draws on GTDB (CC BY-SA 4.0) and SILVA (CC BY 4.0) | Attribution, to the training sets and to Greengenes2. GTDB's share-alike term reaches taxonomic datasets you derive from them and distribute. |
| UNITE general FASTA release (10.0, 19.02.2025) | ITS reference sequences with taxonomy, for fungi and for all eukaryotes | CC BY 4.0 | Attribution, in the form UNITE specifies (below). |
| PR2, the Protist Ribosomal Reference database (5.1.1) | 18S rRNA reference sequences with taxonomy, for eukaryotes and protists in particular | MIT | Keep the copyright and permission notice with any copy you distribute. |
| UniProtKB/Swiss-Prot, UniRef50 and UniRef90 (2026_03), as DIAMOND databases | Reviewed protein sequences, and UniProt sequences clustered at 50% and 90% identity | CC BY 4.0 | Attribution. UniProt requires its copyright statement to accompany every copy, and each hosted copy carries it. |
| mOTUs database (3.1.0), for mOTUs | Marker gene database for taxonomic profiling of metagenomes | CC BY 4.0 | Attribution. |
| AMRFinderPlus database (2026-08-07.1), for AMRFinderPlus | Antimicrobial resistance, stress response and virulence genes and protein models, curated by NCBI | Public domain (U.S. Government work); two of its models are TIGRFAMs models under CC BY-SA 4.0 | NCBI asks for attribution. See the caveat below on individual records. |
| NCBIfam protein family HMMs (20.0) | The protein family models NCBI uses to annotate prokaryotic genomes, including the TIGRFAMs families | CC BY-SA 4.0, for the TIGRFAMs models it contains; NCBI places no restriction on the rest | Attribution, and share-alike for the TIGRFAMs models. |
| TIGRFAMs protein family HMMs (15.0) | Protein family models for prokaryotic genome annotation | CC BY-SA 4.0 | Attribution, by citing the TIGRFAMs papers, and share-alike. |
Attribution statements
If you publish results derived from these databases, these are the credits their licenses call for:
- GTDB — "Taxonomy from the Genome Taxonomy Database (GTDB), used under CC BY-SA 4.0."
- sourmash GTDB sketches — the GTDB statement above applies, since the sketches are derived from GTDB.
- NCBI-derived indexes (Kraken2, Kaiju) — "Built from NCBI RefSeq."
- eggNOG — "Functional annotation from the eggNOG database, used under CC BY 3.0." The same statement covers the taxonomy-scoped sequence databases, which hold eggNOG's own sequences.
- BUSCO — "BUSCO lineage datasets (odb12), used under CC BY-ND 4.0." The publisher requires any publication or product that uses them to cite https://doi.org/10.1093/nar/gkae987.
- CheckM — "CheckM reference data by Donovan Parks, used under CC BY 4.0" (https://doi.org/10.5281/zenodo.7401545).
- Pfam, Rfam and Dfam — no attribution is required. Citing the database you used is customary, and Rfam asks for it.
- Greengenes2 training sets — "Greengenes2 2024.09 training sets for DADA2, used under CC BY 4.0" (https://doi.org/10.5281/zenodo.14169078), with a citation of Greengenes2 itself: McDonald et al., Nature Biotechnology, 2023.
- UNITE — "Abarenkov, Kessy; Zirk, Allan; Piirmann, Timo; Pöhönen, Raivo; Ivanov, Filipp; Nilsson, R. Henrik; Kõljalg, Urmas (2025): UNITE general FASTA release for Fungi. Version 19.02.2025. UNITE Community. https://doi.org/10.15156/BIO/3301229", or, for the all-eukaryote set, the same citation with https://doi.org/10.15156/BIO/3301231.
- PR2 — "PR2, Copyright 2012-2021 D. Vaulot, L. Guillou, J. del Campo, F. Mahé and others, used under the MIT License."
- UniProt — "Protein sequences from UniProt release 2026_03 (the UniProt Consortium), used under CC BY 4.0."
- sourmash GTDB sketches, release 226 — the GTDB statement above applies.
- mOTUs — "mOTUs database 3.1.0, used under CC BY 4.0", citing https://doi.org/10.1186/s40168-022-01410-z.
- AMRFinderPlus — "AMRFinderPlus database, National Center for Biotechnology Information", citing Feldgarden et al., Scientific Reports 11:12728, 2021.
- NCBIfam — "NCBIfam release 20.0, National Center for Biotechnology Information", citing Li, O'Neill et al., Nucleic Acids Research 49:D1020, 2021.
- TIGRFAMs — "TIGRFAMs 15.0, used under CC BY-SA 4.0", citing the four TIGRFAMs papers NCBI lists in its licence statement. The same credit is owed for the TIGRFAMs models inside NCBIfam, CheckM and AMRFinderPlus.
Share-alike is the obligation most easily missed. GTDB and the sourmash sketches built from it are licensed CC BY-SA 4.0, and so are the TIGRFAMs models, whether used on their own or inside NCBIfam, CheckM and AMRFinderPlus; the Greengenes2 training sets carry GTDB's taxonomy. If you distribute a dataset you built from them — a taxonomy table, a curated set of assignments, a model library — that dataset must carry the same license. Using them to produce a result you keep to yourself, or a figure in a paper, does not trigger this.
No-derivatives applies to the BUSCO datasets. You may use them, and share them exactly as published, but you may not distribute a modified version — a trimmed, merged or edited lineage dataset. Results produced with them are not derivatives of the datasets and are yours to publish, with the citation above.
A caveat on the NCBI-derived databases
NCBI states that it places no restrictions on the use or distribution of the data in its molecular databases. It also states that some submitters of the original data, or the country of origin of that data, may claim patent, copyright or other intellectual property rights, and that NCBI cannot assess or grant permission regarding those claims.
This applies to the Kraken2 and Kaiju indexes, to the AMRFinderPlus database and to NCBIfam. UniProt carries a similar notice that some of its data may be covered by patents or other rights.
Ploid is in the same position: we can host and serve these databases on NCBI's terms, and we cannot warrant that no individual record within them is encumbered by a third-party claim. If your use depends on that assurance for particular sequences, verify it against the source records.
Your obligations
Ploid does not hold data licenses on your behalf. Making a database readable by your workflow runs is not a grant of rights in it, and it does not transfer any license to you. If you publish results derived from these sources, the attribution and licensing obligations above are yours to meet.
Contact
Questions about the databases available on your deployment, or a request for one not listed here, can be sent to legal@ploid.ai.