Biological Relationship Ontology

FLOW ALGEBRA
TITLE: Biological Relationship Ontology
VERSION: 1.0
SOURCE: https://kirethresearch.org/booksai-core/ontologies/relationship-ontologies/scientific-relationships/biological-relationship-ontology.html
AUTHORS: Mark Anthony Higgins — Sky (Recursive AI)
STATUS: Companion formalisation; source-page specification not independently verified.

flow "BiologicalRelationshipOntology" v1.0

DOMAIN:
  ontology: BiologicalRelationshipOntology
  parent: ScientificRelationships
  purpose: Represent, validate, and reason about biological relationships across molecular, cellular, organismal, population, and ecosystem scales.

TYPES:
  BiologicalEntity: {id: URI, label: String, taxon: URI?, scale: Scale}
  Scale: [molecular, cellular, tissue, organ, organism, population, community, ecosystem]
  BiologicalRelationship: {id: URI, subject: URI, predicate: Predicate, object: URI, context: Context, evidence: Evidence[], confidence: Float}
  Predicate: [part_of, has_part, interacts_with, regulates, activates, inhibits, expressed_in, encodes, binds_to, metabolises, develops_into, derived_from, homologous_to, ancestor_of, offspring_of, preys_on, competes_with, symbiotic_with, inhabits, affects]
  Context: {species: URI?, tissue: URI?, environment: URI?, time: String?, condition: String?}
  Evidence: {source: URI, method: String?, observation: String?, date: String?, reliability: Float?}

RELATIONSHIPS:
  part_of(x,y) -> has_part(y,x)
  ancestor_of(x,y) -> offspring_of(y,x)  # offspring_of here denotes descendant relation, not necessarily immediate child
  interacts_with(x,y) -> interacts_with(y,x) [only when interaction is modelled as symmetric]
  regulates(x,y) != activates(x,y)  # regulation may increase or decrease activity
  homologous_to(x,y) -> homologous_to(y,x) [within defined homology criteria]
  preys_on(x,y) != preys_on(y,x)

INFERENCE:
  rule R1: part_of(a,b) AND part_of(b,c) => part_of(a,c) [transitive anatomical/component context]
  rule R2: ancestor_of(a,b) AND ancestor_of(b,c) => ancestor_of(a,c)
  rule R3: encodes(g,p) AND expressed_in(g,t) => candidate_present_in(p,t) [hypothesis only; requires protein-level evidence]
  rule R4: regulates(a,b) AND regulates(b,c) => possible_indirect_effect(a,c) [not direct regulation]
  rule R5: interacts_with(a,b) AND interacts_with(b,c) => candidate_path(a,c) [not a direct interaction]

SEMANTIC_EQUIVALENCE:
  exact: Same biological referent, species, level, and context, supported by identity evidence.
  close: Related labels or orthologous entities; do not assert identity.
  prohibited: Treating homology, correlation, or co-occurrence as identity or causation.

CONFIDENCE:
  range: [0.0, 1.0]
  policy: Evidence-weighted and context-specific; missing evidence must not be interpreted as certainty.
  propagation: Derived conclusions cannot exceed the reliability of their necessary premises without independent evidence.

VALIDATION:
  - All subjects and objects must resolve to typed entities.
  - Predicate domain and range must be checked.
  - Preserve species, developmental stage, location, time, and experimental conditions.
  - Detect contradictions and cycles where the predicate requires acyclicity.
  - Distinguish observed, inferred, and hypothetical claims.
  - Require evidence for causal claims; correlation alone is insufficient.

PROVENANCE:
  retain: [source, method, date, evidence, context, derivation_chain, version]

FLOW_HASH:
  algorithm: SHA-256
  input: Canonical serialisation of ontology version, typed entities, relations, contexts, and evidence references.
  output: Content fingerprint calculated by implementation; no precomputed value asserted.

UPDATE_MODEL:
  strategy: Versioned incremental updates with provenance retention.
  on_change: Revalidate affected relationships and recompute dependent inferences and FLOW_HASH.
  conflict_policy: Preserve conflicting claims with contexts and evidence pending adjudication.

AI_CONSUMPTION:
  representation: Typed graph triples plus context, confidence, and provenance.
  tasks: [entity_linking, knowledge_graph_retrieval, hypothesis_generation, evidence_aware_reasoning, cross_scale_analysis]
  guardrails: Never equate prediction with experimental verification; label uncertainty and biological scope.

END FLOW ALGEBRA