FANTASIA Lite V1 Lookup Bundle for Protein Function Annotation

This Zenodo record provides the required lookup bundle for running FANTASIA Lite V1, a lightweight, standalone version of the FANTASIA protein function annotation workflow. FANTASIA Lite V1 performs Gene Ontology (GO) annotation of protein sequences using protein language model embeddings and nearest-neighbor transfer, without requiring PostgreSQL, RabbitMQ, or database-backed orchestration. The software repository is distributed separately through GitHub, while this record provides the flat-file lookup resources required for annotation runs. This lookup bundle must be downloaded and placed in data/lookup/ before running the Lite-V1 pipeline. ContentsThe bundle source is generated from UniProt November 2025 release with experimental annotations dump (https://doi.org/10.5281/zenodo.17720428) and contains contains: lookup_table.npz: precomputed reference embeddings used for nearest-neighbor lookup annotations.json: GO annotation mappings for the reference proteins accessions.json: accession and identifier mappings for the reference set ScopeThe lookup table was generated from the UniProt November 2025 release and includes only proteins with experimental evidence, enabling high-confidence annotation transfer. Reference embedding coverage in the lookup bundle includes: ESM-2 ProstT5 ProtT5 Ankh3-Large ESM3c The current built-in Lite V1 embedding CLI directly supports: ProtT5 Ankh3 So the lookup bundle is broader than the current built-in embedding interface, reflecting the separation between embedding generation and lookup in Lite V1. Intended UseThis record is intended to support: local use of FANTASIA Lite V1 reproducible protein annotation workflows integration of Lite into external bioinformatics pipelines benchmarking and comparison against full FANTASIA workflows Software RepositoryGitHub repository:https://github.com/CBBIO/FANTASIA-Lite Lite V1 branch:FANTASIA-Lite-V1 Installation NoteTo use this bundle with FANTASIA Lite V1: clone the GitHub repository checkout the FANTASIA-Lite-V1 branch place the extracted files in data/lookup/ run the pipeline, for example: ./scripts/minimal_pipeline.sh your_proteins.fa