PASS

PASS provides homology-based protein annotation by maintaining a database of homologous clusters from representative bacterial, archaeal, eukaryal, and viral genomes and integrating machine-learning queries of UniprotKB to incorporate experimental verification of protein functions.

Key Features:

• Extensive homologous-cluster database: Contains over eight million protein sequences compiled from representative genomes of bacteria, archaea, eukarya, and viruses.
• Machine-learning verification: Periodically queries UniprotKB with a machine-learning component to detect experimentally verified protein functions and align annotations with those findings.
• Sequence-query results: Sequence-query functionality returns FASTA headers for sequences within the best-matching homologous cluster.
• Sequence similarity network generation: Generates sequence similarity networks from returned cluster sequences to visualize relationships among proteins.
• Expert-guided annotation selection: Supports selection of annotations from cluster sequences as an alternative to automated assignment.

Scientific Applications:

• Protein function prediction: Improves functional inference by combining homology clustering with experimental verification from UniprotKB.
• Evolutionary studies: Enables comparative analyses of homologous clusters across bacteria, archaea, eukarya, and viruses to study protein evolution.
• Functional genomics: Facilitates annotation of proteins in genome-scale datasets using cluster-based homology and verification status.
• Therapeutic development support: Provides more reliable annotations that can inform target selection and functional interpretation in therapeutic research.

Methodology:

Assembly of large-scale homologous clusters from representative genomes; periodic machine-learning-driven queries of UniprotKB to identify experimentally verified functions and update annotations; generation of sequence similarity networks and provision of cluster-based annotation selection.