Schema

Authoritative source: src/proteintraitsmech/schema/proteintraitsmech.yaml. This page is a plain-English guide; the LinkML file is the machine-readable contract.

← back to index

Root class — ProteinTraitRecord

One YAML file per record. Required slots: identifier, label, trait_axis. Recommended: trait_category, definition, definition_source, mapping_status.

Slot Range Notes
identifier CURIE Prefer source-anchored (e.g. PROSITE:PS00001, TED:AF-…-TED03, UniProtKB:P25888); use proteintraitsmech: for curator-minted
label string Human-readable trait label
definition string Text definition (folded scalar)
definition_source string Release stamp or CURIE-style pointer to the source
synonyms list of TraitSynonym With synonym_type: EXACT / NARROW / BROAD / RELATED
parent_traits list of CURIE rdfs:subClassOf parents
xrefs list of CURIE Cross-references to other ontologies / databases
sequence_pattern string Raw PROSITE / ELM / regex syntax (SEQUENCE-axis records only)
canonical_examples list of CanonicalExample UniProtKB accession + taxon + note
trait_axis TraitAxisEnum See below
trait_category ProteinTraitCategoryEnum See below
term_kind TermKindEnum CLASS / DATATYPE_PROPERTY / OBJECT_PROPERTY / ANNOTATION_PROPERTY
mapping_status MappingStatusEnum SEEDED / PROPOSED / REVIEWED / DEPRECATED
evidence list of EvidenceItem Optional entry-level citation list
causal_graphs list of CausalGraph Evidence-backed mechanism graphs
curation_history list of CurationEvent Append-only audit trail

TraitAxisEnum

Value Meaning
SEQUENCE Trait defined on the linear amino-acid sequence
STRUCTURE Trait defined on the 3D structure (fold, contact, geometry)
SEQUENCE_STRUCTURE Trait meaningful in both axes — transmembrane spans, coiled coils, structural tandem repeats
FUNCTION Entry-level (non-localised) trait — enzymatic activity, binding capacity, cofactor requirement, localisation, environmental response, interaction partner

ProteinTraitCategoryEnum

SEQUENCE

Basic segments: SEQ_MOTIF, SEQ_SIGNAL_PEPTIDE, SEQ_TRANSIT_PEPTIDE, SEQ_PROPEPTIDE, SEQ_INITIATOR_METHIONINE, SEQ_MATURE_CHAIN, SEQ_NONSTANDARD_RESIDUE, SEQ_CLEAVAGE_SITE, SEQ_LOW_COMPLEXITY, SEQ_DISORDER, SEQ_REPEAT, SEQ_COMPOSITION, SEQ_CONSERVATION, SEQ_EPITOPE.

PTM site subtypes (specific — prefer these over the generic bucket): SEQ_MODIFIED_RESIDUE, SEQ_GLYCOSYLATION_SITE, SEQ_LIPIDATION_SITE, SEQ_CROSSLINK_SITE.

Generic PTM bucket (use when the modification class is unknown or heterogeneous): SEQ_PTM_SITE.

Fallback: SEQ_OTHER.

STRUCTURE

CATH/SCOP hierarchy (top-down): STRUCT_CLASS, STRUCT_ARCHITECTURE, STRUCT_TOPOLOGY, STRUCT_HOMOLOGOUS_SUPERFAMILY, STRUCT_FOLD, STRUCT_DOMAIN.

Local geometry: STRUCT_SECONDARY, STRUCT_QUATERNARY, STRUCT_INTERFACE, STRUCT_ACTIVE_SITE, STRUCT_BINDING_SITE, STRUCT_ALLOSTERIC_SITE, STRUCT_DISULFIDE, STRUCT_METAL_SITE, STRUCT_CAVITY, STRUCT_SYMMETRY, STRUCT_DYNAMICS, STRUCT_STABILITY, STRUCT_SURFACE, STRUCT_OTHER.

SEQUENCE_STRUCTURE (mixed)

MIXED_TRANSMEMBRANE, MIXED_COILED_COIL, MIXED_STRUCTURAL_REPEAT, MIXED_OTHER.

FUNCTION

Category Grounding
FUNC_ENZYMATIC_ACTIVITY EC, Rhea, GO MF *activity, participating ChEBIs
FUNC_BINDING_CAPACITY GO MF * binding, UniProt KW
FUNC_COFACTOR_REQUIREMENT ChEBI + UniProt CC COFACTOR
FUNC_LOCALIZATION UniProt SubCell, GO CC
FUNC_ENVIRONMENTAL_RESPONSE GO BP response_to_*, UniProt CC INDUCTION keyword scan
FUNC_INTERACTION_PARTNER UniProt CC SUBUNIT, IntAct / BioGRID / STRING

Administrative

UPPER (organisational parent), OTHER.

Axis / category pairing

ProteinTraitRecord carries LinkML rules that enforce the category prefix against the axis:

Category prefix Required axis
SEQ_* SEQUENCE
STRUCT_* STRUCTURE
MIXED_* SEQUENCE_STRUCTURE
FUNC_* FUNCTION
UPPER / OTHER unconstrained

just validate-all will reject a mismatched pair.

FUNCTION vs localised STRUCTURE

FUNCTION and STRUCTURE records for the same protein are complementary, not duplicative. A UniProt entry with an ATP-binding pocket at residues 45–52 emits both a STRUCT_BINDING_SITE record (localised: where ATP binds) and a FUNC_BINDING_CAPACITY record (entry-level: that the protein binds ATP). Likewise, a catalytic residue emits both STRUCT_ACTIVE_SITE and FUNC_ENZYMATIC_ACTIVITY. The two answer different questions and are meant to coexist.

CausalGraph / CausalNode / CausalEdge

Optional inline mechanism graph attached to a record. Every CausalEdge requires ≥1 EvidenceItem — mechanism assertions are curator-generated and must cite their support.

CausalNodeTypeEnum covers PROTEIN, DOMAIN, MOTIF, RESIDUE, PTM, LIGAND, NUCLEIC_ACID, CHEMICAL, PATHWAY, MOLECULAR_FUNCTION, BIOLOGICAL_PROCESS, CELLULAR_LOCALIZATION, PHENOTYPE, DISEASE, TRAIT, STATE, QUALITY, ENVIRONMENTAL_FACTOR, EXPERIMENTAL_FACTOR, OTHER.

Prefixes declared in the schema

Ontologies: PR, GO, PATO, UBERON, CL, CHEBI, RHEA, SO, UO, IAO, RO, OBI, MOD, MONDO, HP, NCBITaxon. Databases: UniProtKB, Pfam, InterPro, PROSITE, SMART, CATH, SCOP, MEROPS, HAMAP, EC, KEGG, Reactome, NCBIGene, HGNC, TED, AlphaFoldDB, valuesets. Provenance: PMID, DOI, dcterms, skos, biolink.

← back to index