AI Biomedical Reasoning Ontology

AI SEARCH

FLOW ALGEBRA — AI BIOMEDICAL REASONING ONTOLOGY
Semantic Library
Version: 4.0
Canonical Page:
https://kirethresearch.org/booksai-core/ontologies/biological-ontologies/ai-biology/ai-biomedical-reasoning-ontology.html

FLOW AI_BIOMEDICAL_REASONING_ONTOLOGY {

  META {
    DOMAIN = "Biological Ontologies"
    SUBDOMAIN = "AI Biology"
    ONTOLOGY = "AI Biomedical Reasoning Ontology"
    PURPOSE = "Formalise entities, relationships, evidence structures and computational reasoning processes used by artificial intelligence to integrate and reason over biomedical knowledge."
    LOCALE = "en-GB"
  }

  ENTITY_TYPES {
    AI_SYSTEM
    AI_MODEL
    REASONING_MODEL
    MACHINE_LEARNING_MODEL
    FOUNDATION_MODEL
    BIOMEDICAL_ENTITY
    GENE
    GENETIC_VARIANT
    PROTEIN
    MOLECULE
    DRUG
    DRUG_CANDIDATE
    BIOMARKER
    PATHWAY
    BIOLOGICAL_PROCESS
    CELL
    TISSUE
    ORGAN
    ORGANISM
    PHENOTYPE
    TRAIT
    DISEASE
    CONDITION
    SYMPTOM
    SIGN
    DIAGNOSIS
    INTERVENTION
    TREATMENT
    TEST
    LABORATORY_RESULT
    OBSERVATION
    CLINICAL_FINDING
    PATIENT_STATE
    POPULATION
    COHORT
    STUDY
    EXPERIMENT
    DATASET
    LITERATURE_SOURCE
    CLAIM
    HYPOTHESIS
    EVIDENCE
    SUPPORTING_EVIDENCE
    CONTRADICTORY_EVIDENCE
    CAUSAL_RELATION
    ASSOCIATION
    MECHANISM
    RISK_FACTOR
    OUTCOME
    KNOWLEDGE_GRAPH
    BIOLOGICAL_ONTOLOGY
    BIOMEDICAL_ONTOLOGY
    CONCEPT
    ENTITY_REPRESENTATION
    INFERENCE
    EXPLANATION
    CONFIDENCE_SCORE
    UNCERTAINTY
    PROVENANCE_RECORD
  }

  RELATIONSHIPS {
    IS_A
    PART_OF
    LOCATED_IN
    ENCODES
    EXPRESSED_IN
    INTERACTS_WITH
    BINDS_TO
    REGULATES
    ACTIVATES
    INHIBITS
    PARTICIPATES_IN
    ASSOCIATED_WITH
    CAUSES
    CONTRIBUTES_TO
    INCREASES_RISK_OF
    DECREASES_RISK_OF
    HAS_PHENOTYPE
    HAS_SYMPTOM
    HAS_SIGN
    INDICATES
    DIAGNOSED_BY
    MEASURED_BY
    TREATED_BY
    TARGETED_BY
    RESPONDS_TO
    CONTRAINDICATES
    PRECEDES
    FOLLOWS
    CO_OCCURS_WITH
    SUPPORTS
    CONTRADICTS
    DERIVED_FROM
    INFERRED_FROM
    EXPLAINED_BY
    PREDICTS
    CLASSIFIES
    PRIORITISES
    HAS_EVIDENCE
    HAS_CONFIDENCE
    HAS_UNCERTAINTY
    HAS_PROVENANCE
    MAPS_TO
    SEMANTICALLY_EQUIVALENT_TO
  }

  INPUTS {
    BIOMEDICAL_KNOWLEDGE
    BIOLOGICAL_DATA
    GENOMIC_DATA
    PROTEOMIC_DATA
    MOLECULAR_DATA
    PHENOTYPE_DATA
    DISEASE_DATA
    DRUG_DATA
    CLINICAL_FINDINGS
    LABORATORY_DATA
    POPULATION_DATA
    EXPERIMENTAL_DATA
    CLINICAL_STUDIES
    SCIENTIFIC_LITERATURE
    KNOWLEDGE_GRAPHS
    ONTOLOGY_TERMS
    EXISTING_CLAIMS
    EXPERT_ANNOTATIONS
  }

  NORMALISE {
    STANDARDISE_IDENTIFIERS
    RESOLVE_BIOMEDICAL_ENTITIES
    NORMALISE_TERMINOLOGY
    MAP_GENE_IDENTIFIERS
    MAP_PROTEIN_IDENTIFIERS
    MAP_DISEASE_TERMS
    MAP_DRUG_IDENTIFIERS
    MAP_PHENOTYPE_TERMS
    ALIGN_ONTOLOGY_TERMS
    NORMALISE_EVIDENCE_TYPES
    REMOVE_DUPLICATES
    VALIDATE_METADATA
    PRESERVE_PROVENANCE
  }

  PROCESS BIOMEDICAL_REASONING_PIPELINE {
    INPUT -> ENTITY_EXTRACTION
    ENTITY_EXTRACTION -> ENTITY_RESOLUTION
    ENTITY_RESOLUTION -> RELATION_EXTRACTION
    RELATION_EXTRACTION -> KNOWLEDGE_INTEGRATION
    KNOWLEDGE_INTEGRATION -> EVIDENCE_ALIGNMENT
    EVIDENCE_ALIGNMENT -> INFERENCE_GENERATION
    INFERENCE_GENERATION -> CONSISTENCY_CHECK
    CONSISTENCY_CHECK -> CONFIDENCE_ESTIMATION
    CONFIDENCE_ESTIMATION -> EXPLANATION_GENERATION
    EXPLANATION_GENERATION -> OUTPUT
  }

  PROCESS EVIDENCE_REASONING {
    CLAIM + SUPPORTING_EVIDENCE + CONTRADICTORY_EVIDENCE
      -> EVIDENCE_ASSESSMENT

    EVIDENCE_ASSESSMENT
      -> SOURCE_QUALITY
      -> RELEVANCE
      -> CONSISTENCY
      -> REPLICATION
      -> DIRECTNESS

    SOURCE_QUALITY
    + RELEVANCE
    + CONSISTENCY
    + REPLICATION
    + DIRECTNESS
      -> CONFIDENCE_SCORE
  }

  PROCESS CAUSAL_REASONING {
    OBSERVATION + ASSOCIATION + TEMPORAL_RELATION + MECHANISTIC_EVIDENCE
      -> CAUSAL_HYPOTHESIS

    CAUSAL_HYPOTHESIS
      -> ALTERNATIVE_EXPLANATION_CHECK
      -> CONFOUNDING_CHECK
      -> EVIDENCE_INTEGRATION
      -> CAUSAL_CONFIDENCE
  }

  PROCESS DIAGNOSTIC_REASONING {
    SYMPTOM + SIGN + LABORATORY_RESULT + CLINICAL_FINDING
      -> CANDIDATE_EXPLANATIONS

    CANDIDATE_EXPLANATIONS
      -> EVIDENCE_MATCHING
      -> CONTRADICTION_CHECK
      -> DIFFERENTIAL_PRIORITISATION
      -> CONFIDENCE_ESTIMATION

    CONFIDENCE_ESTIMATION
      -> EXPLANATION_PATH
  }

  PROCESS THERAPEUTIC_REASONING {
    DISEASE + TARGET + DRUG + BIOMARKER + EVIDENCE
      -> TREATMENT_HYPOTHESIS

    TREATMENT_HYPOTHESIS
      -> MECHANISM_CHECK
      -> RESPONSE_EVIDENCE_CHECK
      -> SAFETY_CONTEXT_CHECK
      -> CONTRAINDICATION_CHECK
      -> CONFIDENCE_SCORE
  }

  PROCESS MECHANISTIC_REASONING {
    GENE + PROTEIN + PATHWAY + CELL + PHENOTYPE + DISEASE
      -> MECHANISTIC_GRAPH

    MECHANISTIC_GRAPH
      -> PATH_DISCOVERY
      -> RELATION_INFERENCE
      -> MECHANISM_HYPOTHESIS
      -> EVIDENCE_VALIDATION
  }

  PROCESS KNOWLEDGE_GRAPH_REASONING {
    BIOMEDICAL_ENTITY + RELATIONSHIPS + EVIDENCE
      -> BIOMEDICAL_KNOWLEDGE_GRAPH

    BIOMEDICAL_KNOWLEDGE_GRAPH
      -> ENTITY_RESOLUTION
      -> LINK_PREDICTION
      -> PATH_REASONING
      -> RELATION_INFERENCE
      -> EXPLANATION_PATH
  }

  AI_TASKS {
    BIOMEDICAL_ENTITY_RECOGNITION
    ENTITY_LINKING
    RELATION_EXTRACTION
    KNOWLEDGE_INTEGRATION
    EVIDENCE_RETRIEVAL
    EVIDENCE_SYNTHESIS
    HYPOTHESIS_GENERATION
    CAUSAL_REASONING
    MECHANISTIC_REASONING
    DIAGNOSTIC_REASONING
    THERAPEUTIC_REASONING
    DRUG_TARGET_REASONING
    BIOMARKER_REASONING
    RISK_REASONING
    KNOWLEDGE_GRAPH_COMPLETION
    CONTRADICTION_DETECTION
    UNCERTAINTY_ESTIMATION
    EXPLANATION_GENERATION
    BIOMEDICAL_QUESTION_ANSWERING
  }

  RULES {
    IF CLAIM_EXISTS
    THEN LINK {
      CLAIM
      EVIDENCE
      CONFIDENCE_SCORE
      PROVENANCE_RECORD
    }

    IF RELATION == CAUSES
    THEN REQUIRE {
      CAUSAL_EVIDENCE
      ALTERNATIVE_EXPLANATION_CHECK
      PROVENANCE_RECORD
    }

    IF ASSOCIATION_EXISTS
    THEN DO_NOT_AUTOMATICALLY_INFER_CAUSATION

    IF AI_MODEL_GENERATES_INFERENCE
    THEN LABEL_AS_INFERRED_UNLESS_DIRECTLY_OBSERVED

    IF CONTRADICTORY_EVIDENCE_EXISTS
    THEN PRESERVE_CONTRADICTION

    IF EVIDENCE_IS_POPULATION_SPECIFIC
    THEN PRESERVE_POPULATION_CONTEXT

    IF EVIDENCE_IS_PRECLINICAL
    THEN DO_NOT_ASSERT_CLINICAL_VALIDATION

    IF TREATMENT_RESPONSE_IS_OBSERVED
    THEN DO_NOT_GENERALISE_WITHOUT_POPULATION_AND_CONTEXT_CHECK

    IF SEMANTIC_IDENTIFIERS_DIFFER
    AND BIOMEDICAL_REFERENT_IS_SAME
    THEN APPLY SEMANTIC_EQUIVALENCE
  }

  CONFIDENCE {
    CONFIDENCE_SCORE =
      WEIGHT(
        SOURCE_QUALITY,
        EVIDENCE_DIRECTNESS,
        REPLICATION,
        CONSISTENCY,
        BIOLOGICAL_PLAUSIBILITY,
        MODEL_VALIDATION,
        POPULATION_RELEVANCE,
        PROVENANCE_COMPLETENESS
      )

    LEVELS {
      HIGH
      MODERATE
      LOW
      INSUFFICIENT
    }

    IF CONFLICTING_EVIDENCE
    THEN REDUCE CONFIDENCE_SCORE

    IF INDEPENDENT_REPLICATION
    THEN INCREASE CONFIDENCE_SCORE

    IF DIRECT_EXPERIMENTAL_SUPPORT
    THEN INCREASE CONFIDENCE_SCORE

    IF EVIDENCE_CHAIN_CONTAINS_UNCERTAIN_LINK
    THEN PROPAGATE_UNCERTAINTY
  }

  SEMANTIC_EQUIVALENCE {
    EQUIVALENCE_KEY =
      NORMALISE(
        ENTITY_TYPE,
        CANONICAL_IDENTIFIER,
        ONTOLOGY_TERM,
        BIOLOGICAL_CONTEXT,
        CLINICAL_CONTEXT
      )

    IF EQUIVALENCE_KEY(A) == EQUIVALENCE_KEY(B)
    THEN A SEMANTICALLY_EQUIVALENT_TO B

    ELSE PRESERVE_DISTINCT_ENTITIES
  }

  REASONING_CONSTRAINTS {
    ASSOCIATION != CAUSATION
    CORRELATION != MECHANISM
    PREDICTION != OBSERVATION
    INFERENCE != ESTABLISHED_FACT
    PRECLINICAL_EVIDENCE != CLINICAL_VALIDATION
    MODEL_CONFIDENCE != BIOMEDICAL_CERTAINTY
    ABSENCE_OF_EVIDENCE != EVIDENCE_OF_ABSENCE
    POPULATION_CONTEXT_MUST_BE_PRESERVED
    TEMPORAL_CONTEXT_MUST_BE_PRESERVED
    CONTRADICTORY_EVIDENCE_MUST_NOT_BE_DISCARDED
    PROVENANCE_MUST_BE_PRESERVED
    UNCERTAINTY_MUST_NOT_BE_DISCARDED
  }

  VALIDATION {
    CHECK_ENTITY_TYPES
    CHECK_RELATION_DOMAINS_AND_RANGES
    CHECK_IDENTIFIER_RESOLUTION
    CHECK_ONTOLOGY_ALIGNMENT
    CHECK_EVIDENCE_LINKS
    CHECK_CAUSAL_ASSERTIONS
    CHECK_CONTRADICTIONS
    CHECK_POPULATION_CONTEXT
    CHECK_TEMPORAL_CONTEXT
    CHECK_PROVENANCE
    CHECK_CONFIDENCE
    CHECK_SEMANTIC_CONSISTENCY
    CHECK_DUPLICATE_ENTITIES
  }

  OUTPUTS {
    NORMALISED_BIOMEDICAL_ENTITIES
    BIOMEDICAL_RELATIONSHIPS
    INFERRED_RELATIONSHIPS
    CAUSAL_HYPOTHESES
    MECHANISTIC_HYPOTHESES
    DIAGNOSTIC_HYPOTHESES
    THERAPEUTIC_HYPOTHESES
    EVIDENCE_ASSESSMENTS
    KNOWLEDGE_GRAPH_EDGES
    CONTRADICTION_RECORDS
    CONFIDENCE_SCORES
    UNCERTAINTY_RECORDS
    EVIDENCE_LINKS
    PROVENANCE_RECORDS
    EXPLANATION_PATHS
  }

  UPDATE_MODEL {
    ON NEW_BIOMEDICAL_DATA {
      NORMALISE
      ENTITY_RESOLUTION
      UPDATE_RELATIONSHIPS
      RECOMPUTE_RELEVANT_INFERENCES
      UPDATE_CONFIDENCE
      PRESERVE_VERSION_HISTORY
    }

    ON NEW_EVIDENCE {
      LINK_EVIDENCE
      REASSESS_SUPPORT
      REASSESS_CONTRADICTIONS
      REASSESS_CAUSAL_CLAIMS
      UPDATE_CONFIDENCE
      RECORD_PROVENANCE
    }

    ON ONTOLOGY_REVISION {
      REMAP_TERMS
      CHECK_SEMANTIC_EQUIVALENCE
      IDENTIFY_DEPRECATED_TERMS
      MIGRATE_RELATIONSHIPS
      PRESERVE_PRIOR_VERSION_REFERENCES
    }
  }

  FLOW_HASH {
    HASH_INPUT =
      CANONICALISE(
        ENTITY_TYPES,
        RELATIONSHIPS,
        RULES,
        CONFIDENCE,
        SEMANTIC_EQUIVALENCE,
        REASONING_CONSTRAINTS,
        UPDATE_MODEL
      )

    FLOW_HASH = SHA256(HASH_INPUT)
  }

  AI_CONSUMPTION {
    INDEX {
      ONTOLOGY
      ENTITY_TYPES
      RELATIONSHIPS
      AI_TASKS
      RULES
      OUTPUTS
    }

    RETRIEVE {
      BY_ENTITY
      BY_DISEASE
      BY_GENE
      BY_PROTEIN
      BY_DRUG
      BY_PHENOTYPE
      BY_RELATION
      BY_CLAIM
      BY_EVIDENCE
      BY_CONFIDENCE
    }

    REASON {
      APPLY_RELATIONSHIPS
      APPLY_RULES
      INTEGRATE_EVIDENCE
      PRESERVE_UNCERTAINTY
      CHECK_CONTRADICTIONS
      TRACE_EVIDENCE
      TRACE_PROVENANCE
      GENERATE_EXPLANATION_PATH
    }

    RESPONSE_POLICY {
      DISTINGUISH_OBSERVED_FROM_INFERRED
      DISTINGUISH_ASSOCIATION_FROM_CAUSATION
      DISTINGUISH_PRECLINICAL_FROM_CLINICAL
      RETURN_CONFIDENCE_WHEN_AVAILABLE
      RETURN_PROVENANCE_WHEN_AVAILABLE
      EXPOSE_CONFLICTING_EVIDENCE
      PRESERVE_CONTEXT
      AVOID_UNSUPPORTED_BIOMEDICAL_INFERENCE
    }
  }

  TERMINAL {
    RESULT =
      VALIDATED(
        SEMANTICALLY_NORMALISED(
          EVIDENCE_LINKED(
            CONFIDENCE_AWARE(
              AI_BIOMEDICAL_REASONING_KNOWLEDGE
            )
          )
        )
      )
  }
}

END FLOW