id: https://ai4curation.io/ai-gene-review/schema/pfam_entry_review name: pfam_entry_review title: Pfam Entry Review Schema description: >- Sidecar LinkML schema for *curating a single Pfam family as its own entry*, stored next to the machine-fetched record at ``interpro/pfam//-review.yaml`` (the fetched ``-metadata.yaml`` is never hand-edited). This is the entry-centric alternative to a flat SSSOM mapping set: instead of one subject->predicate->object row per assertion, a review captures the full curation context for one Pfam family. It records (1) an explicit ``interpro`` block -- the parent InterPro entry, every member Pfam lumped under it, and why a GO mapping at the entry level is or is not viable; and (2) per proposed GO annotation, the relation/aspect/confidence/status PLUS the grounding that distinguishes a real family-function alignment from a family merely *named* after a function: ``supporting_examples`` (characterized SwissProt members that have the function, linked to in-repo gene reviews where they exist) and ``counter_examples`` (members -- sibling families, or the SAME family -- that lack/contradict the function and would be mis-annotated). A proposal whose own family contains functionally distinct members (counter_examples within the family) is recorded with ``status: REJECTED``. GO targets are id/label tuples bound to a GO branch enum so that ``linkml-term-validator`` checks every proposed term id resolves and its label matches the ontology. Pfam/InterPro/UniProt ids are curated id/label tuples and are not ontology-label-validated here. prefixes: linkml: https://w3id.org/linkml/ pfamreview: https://ai4curation.io/ai-gene-review/pfam_entry_review/ Pfam: https://www.ebi.ac.uk/interpro/entry/pfam/ InterPro: https://www.ebi.ac.uk/interpro/entry/InterPro/ UniProt: https://www.uniprot.org/uniprotkb/ GO: http://purl.obolibrary.org/obo/GO_ RO: http://purl.obolibrary.org/obo/RO_ BFO: http://purl.obolibrary.org/obo/BFO_ rdfs: http://www.w3.org/2000/01/rdf-schema# imports: - linkml:types default_prefix: pfamreview default_range: string slots: id: identifier: true description: CURIE of the term (e.g. Pfam:PF14681, GO:0004845). label: slot_uri: rdfs:label description: Human-readable label (validated against the ontology where the slot is enum-bound). # --- top-level entry slots --- pfam_id: range: string required: true pattern: "^PF[0-9]{5}$" description: The Pfam accession this review is about (e.g. PF14681). pfam_name: range: string required: true description: Pfam family short name (e.g. UPRTase). pfam_description: range: string description: Short curator-facing description of the family's function. interpro: range: InterProContext inlined: true required: true description: The parent InterPro entry, its member Pfams, and the entry-level mapping viability. proposed_annotations: range: ProposedAnnotation inlined_as_list: true multivalued: true description: Proposed (or rejected) GO annotations for THIS Pfam family. notes: range: string description: Free-text curator notes. # --- InterProContext slots --- type: range: string description: InterPro entry type (Domain, Family, Homologous_superfamily, Repeat, ...). member_pfams: range: MemberPfam inlined_as_list: true multivalued: true description: All Pfam families that are members of the InterPro entry (including the subject family). go_status: range: InterProGOStatusEnum required: true description: Whether/how the parent InterPro entry currently carries an interpro2go term. mapping_viability: range: MappingViabilityEnum required: true description: >- Whether a function-specific GO mapping is viable at the InterPro entry level. NOT_VIABLE is the typical case: the entry lumps functionally distinct member families, so a specific term cannot sit on the whole entry. viability_reason: range: string description: Why the InterPro-level mapping is (not) viable. # --- MemberPfam slots --- is_subject: range: boolean description: True for the Pfam family this review is about. note: range: string description: Free-text note. # --- ProposedAnnotation slots --- term: range: GOTerm inlined: true required: true description: The proposed GO term (id/label), validated against the GO branch enum. bindings: - binds_value_of: id range: GOTermEnum obligation_level: REQUIRED relation: range: RelationTerm inlined: true required: true description: How the Pfam relates to the GO term (enables for MF, involved_in for BP, part_of for CC). aspect: range: GOAspectEnum required: true description: GO aspect of the proposed term. confidence: range: ConfidenceEnum required: true description: Reviewer confidence in the proposed annotation. status: range: AnnotationStatusEnum required: true description: Curation status of the proposed annotation. rationale: range: string required: true description: >- Reviewer rationale: for a PROPOSED annotation, why the family's HMM tracks the function (not merely shares its name); for a REJECTED one, why it does not. supporting_examples: range: ProteinExample inlined_as_list: true multivalued: true description: Characterized members of THIS family that have the proposed function. counter_examples: range: CounterExample inlined_as_list: true multivalued: true description: >- Members that lack/contradict the function and would be mis-annotated -- in a sibling member family, or (for REJECTED proposals) in the SAME family. # --- ProteinExample / CounterExample shared slots --- accession: range: string required: true pattern: "^UniProt:[A-Z0-9]+$" description: UniProt accession (e.g. UniProt:P0A8F0). protein_name: range: string description: UniProt entry/protein name (e.g. UPP_ECOLI uracil phosphoribosyltransferase). organism: range: string description: Source organism. ec: range: string description: EC number(s), if the member is an enzyme with an assigned activity. reviewed: range: boolean description: True if the UniProt entry is SwissProt-reviewed. gene_review: range: string description: Path to an in-repo gene review for this protein, if one exists (e.g. genes/BACSU/spoIIE/spoIIE-ai-review.yaml). member_pfam: range: string pattern: "^Pfam:PF[0-9]{5}$" description: The member Pfam family this counter-example belongs to (same family for REJECTED, else a sibling). reason: range: string required: true description: Why this protein lacks/contradicts the proposed function. classes: InterProContext: description: The parent InterPro entry, its member Pfams, and the entry-level mapping viability. slots: - id - label - type - member_pfams - go_status - mapping_viability - viability_reason slot_usage: id: pattern: "^InterPro:IPR[0-9]{6}$" required: true MemberPfam: description: A Pfam family that is a member of the InterPro entry. slots: - id - label - is_subject - note slot_usage: id: pattern: "^Pfam:PF[0-9]{5}$" GOTerm: description: A GO term as an id/label pair. slots: - id - label slot_usage: id: pattern: "^GO:[0-9]{7}$" label: required: true RelationTerm: description: The Pfam->GO relation as an id/label pair (RO/BFO predicate). slots: - id - label ProteinExample: description: A characterized member protein that has the proposed function. slots: - accession - protein_name - organism - ec - reviewed - gene_review - note CounterExample: description: A member protein that lacks/contradicts the proposed function (would be mis-annotated). slots: - accession - protein_name - organism - ec - member_pfam - reviewed - gene_review - reason ProposedAnnotation: description: One proposed (or rejected) GO annotation for the Pfam family. slots: - term - relation - aspect - confidence - status - rationale - supporting_examples - counter_examples PfamEntryReview: tree_root: true description: A curated review of a single Pfam family (entry-centric, not a flat mapping row). slots: - pfam_id - pfam_name - pfam_description - interpro - proposed_annotations - notes enums: InterProGOStatusEnum: permissible_values: ABSENT: description: The parent InterPro entry has no interpro2go term. PRESENT_GENERAL: description: The parent InterPro entry has only general (high-level) GO term(s). PRESENT_SPECIFIC: description: The parent InterPro entry already has a specific GO term. MappingViabilityEnum: permissible_values: NOT_VIABLE: description: >- A function-specific GO term cannot be placed on the InterPro entry (it would be wrong for a functionally distinct member); a viable term, if any, belongs at the Pfam level. VIABLE: description: >- A GO term could equally be placed at the InterPro entry level (e.g. the entry is 1 Pfam = 1 entry, or all members share the function); recorded for completeness. GOAspectEnum: permissible_values: molecular_function: biological_process: cellular_component: ConfidenceEnum: permissible_values: HIGH: MEDIUM: LOW: AnnotationStatusEnum: permissible_values: PROPOSED: description: Candidate proposed by review; needs curator / experimental validation. ACCEPTED: description: Confirmed by a curator. REJECTED: description: >- Rejected after member verification -- typically because the Pfam family itself is functionally heterogeneous (see counter_examples in the same family), so the term would over-annotate even at the Pfam level. GOTermEnum: description: Any term in the Gene Ontology (MF, BP, or CC branch). reachable_from: source_nodes: - GO:0003674 - GO:0008150 - GO:0005575 is_direct: false include_self: true relationship_types: - rdfs:subClassOf