Legal and attributions
Publisher
Sequality and this site are published by OneGlobe LLC. Questions about licensing go to michael.harrison@oneglobeit.com.
Open-source tools and model on the search path
These are the open-source components on the shipped search path, with the license each is used under.
| Component | Used for | License |
|---|---|---|
| ESM-2 650M | Protein embeddings for AI similarity retrieval and for grouping references | MIT |
| MMseqs2 | Candidate search and alignment | MIT |
| HMMER | Profile hidden Markov model search and profile building | BSD |
| parasail | Smith-Waterman re-ranking of protein candidates | BSD |
| MAFFT | Multiple sequence alignment and consensus | BSD |
| EMBOSS | Pattern search and sequence utilities | GPL. It runs as its own tool container and is called as a separate program; no product code links it. |
The ESM-2 weights are self-hosted and pinned to a recorded revision. Amazon Nova Lite, the language model the product calls today, is an AWS service used through Amazon Bedrock under AWS's terms; it is not bundled with the product. Research channels that are not on the shipped search path are not listed here.
A complete third-party notice is planned.
Public data
Sequality can search public sequence databases. Each is used under its own terms, and each hit names the database it came from.
- UniProtKB/Swiss-Prot, © The UniProt Consortium, CC BY 4.0
- UniProtKB/Swiss-Prot is licensed under Creative Commons Attribution 4.0 International.
- NCBI GenBank patent sequences (PAT) are used under the NCBI website and data usage policies.
This site
The text of this site describes the product as built at the commit it was generated from. The screenshots show test fixture data: placeholder records and patent numbers, not search results.