What this entity is
Canonical name: Swiss-Prot. 1
Known aliases: Swiss-Prot; reviewed UniProt protein sequences. 1
Stable external identifier: No stable external identifier was reported in this document. 1
Swiss-Prot is the reviewed protein-sequence dataset used to provide enzyme-homology exemplar sequences for the annotation-gap discovery pipeline. 1
Key facts
Annotation-gap discovery
The study retrieved 328 reviewed bacterial sequences for 75 of 84 unique EC numbers through the UniProt REST API. 1
The sequences served as exemplars for DIAMOND blastp homology searches against candidate proteins from 14 Fitness Browser organisms. 1
The DIAMOND v2.1.16 search used --evalue 1e-5, --max-target-seqs 20, --id 25, --query-cover 50, and --outfmt 6. 1
The search identified 154 BLAST hits. 1
High-confidence homology required at least 30% identity, at least 70% coverage, and an e-value at most 1e-10, while medium-confidence homology required at least 25% identity, at least 50% coverage, and an e-value at most 1e-5. 1
BLAST was the strongest single evidence stream: it resolved 70 of 201 gapfilled enzymatic reaction-organism pairs, or 34.8%, whereas the complete evidence pipeline resolved 96 of 201 pairs, or 47.8%. 1
The complete pipeline combined Swiss-Prot homology with fitness evidence, pangenome conservation, pathway evidence, and metabolic-model gapfilling; this integration added 13 percentage points over BLAST alone. 1
The reactions with the most BLAST hits were rxn02185, rxn03436, and rxn15947, which the study characterized as well-described enzymes with broad phylogenetic distribution. 1
Collection-scale literature coverage
The PaperBLAST analysis refines Swiss-Prot’s role from a homology reference to a partially literature-linked protein resource: it contains 110,171 Swiss-Prot proteins associated with 181,916 unique papers, with 1.7 reported papers per protein. 2
PaperBLAST includes approximately 19% of the full SwissProt database, estimated there as approximately 570K reviewed entries as of 2024; this qualifies rather than contradicts the use of Swiss-Prot as a reference, because curated knowledge exists for proteins that PaperBLAST cannot connect to PMC full text. 2
The PaperBLAST coverage analysis supports the importance of provenance when interpreting Swiss-Prot-linked literature: its links are based on PubMed Central full-text mining, so papers behind paywalls may be missed, and a gene mention is not necessarily a functional characterization. 2
Structural fitness integration
The BERDL Data Atlas refines Swiss-Prot’s role as a homology reference by documenting a validated structural-fitness join: FitnessBrowser gene-fitness records were joined to Swiss-Prot best hits using the composite orgId, locusId key, then to AlphaFold entries through sprotAccession and uniprot_accession. 3
That validation found 79,180 genes with a Swiss-Prot best hit, of which 78,753 were represented in AlphaFold; 55,454 genes across 48 organisms had both fitness data and an AlphaFold model. 3 The joined cohort contained 22,303 distinct AlphaFold models. 3
Related pages
Swiss-Prot evidence is connected to the diamond homology-search method and the uniprot protein resource. 1
The dataset contributes to metabolic-model-gapfilling by supplying sequence evidence for candidate genes underlying gapfilled reactions. 1
It also contributes to multi-omics-integration through integration of protein homology with fitness, pangenome, pathway, and metabolic-model evidence. 1
The validated structural-fitness use further connects Swiss-Prot to kescience-alphafold and condition-specific-fitness. 3
See the source summaries: Annotation-Gap Discovery via Phenotype-Fitness-Pangenome-Gapfilling Integration, BERDL Data Atlas — Inventory, Topic Map, and Cross-Reference Synergies, and PaperBLAST Data Explorer — Literature Coverage Bias in Protein Sequence Space. 1,3,2