AI Protein Prediction Ontology

AI SEARCH

FLOW ALGEBRA — AI PROTEIN PREDICTION ONTOLOGY
Semantic Library
Version: 4.0
Canonical Page:
https://kirethresearch.org/booksai-core/ontologies/biological-ontologies/ai-biology/ai-protein-prediction-ontology.html

FLOW AI_PROTEIN_PREDICTION_ONTOLOGY {

  META {
    DOMAIN = "Biological Ontologies"
    SUBDOMAIN = "AI Biology"
    ONTOLOGY = "AI Protein Prediction Ontology"
    PURPOSE = "Formalise entities, relationships and computational reasoning structures connecting artificial intelligence with protein sequence, structure, function, interaction and property prediction."
    LOCALE = "en-GB"
  }

  ENTITY_TYPES {
    AI_SYSTEM
    AI_MODEL
    MACHINE_LEARNING_MODEL
    DEEP_LEARNING_MODEL
    FOUNDATION_MODEL
    PROTEIN_MODEL
    PROTEIN
    PROTEIN_SEQUENCE
    AMINO_ACID
    RESIDUE
    PEPTIDE
    DOMAIN
    MOTIF
    PROTEIN_FAMILY
    PROTEIN_STRUCTURE
    PRIMARY_STRUCTURE
    SECONDARY_STRUCTURE
    TERTIARY_STRUCTURE
    QUATERNARY_STRUCTURE
    STRUCTURAL_REGION
    BINDING_SITE
    ACTIVE_SITE
    INTERFACE
    CONFORMATION
    FOLD
    GENE
    TRANSCRIPT
    LIGAND
    MOLECULE
    DRUG
    RECEPTOR
    ENZYME
    PATHWAY
    CELL
    ORGANISM
    FUNCTION
    BIOLOGICAL_PROCESS
    MOLECULAR_FUNCTION
    CELLULAR_COMPONENT
    INTERACTION
    PROTEIN_PROTEIN_INTERACTION
    PROTEIN_LIGAND_INTERACTION
    MUTATION
    VARIANT
    FEATURE
    EMBEDDING
    REPRESENTATION
    ALIGNMENT
    HOMOLOG
    TEMPLATE
    PREDICTION
    CONFIDENCE_SCORE
    EVIDENCE
    DATASET
    KNOWLEDGE_GRAPH
    BIOLOGICAL_ONTOLOGY
    PROVENANCE_RECORD
  }

  RELATIONSHIPS {
    HAS_SEQUENCE
    CONTAINS_RESIDUE
    HAS_DOMAIN
    HAS_MOTIF
    HAS_STRUCTURE
    HAS_SECONDARY_STRUCTURE
    HAS_TERTIARY_STRUCTURE
    HAS_QUATERNARY_STRUCTURE
    HAS_FOLD
    HAS_BINDING_SITE
    HAS_ACTIVE_SITE
    HAS_INTERFACE
    HAS_CONFORMATION
    ENCODES
    HOMOLOGOUS_TO
    SIMILAR_TO
    ALIGNS_WITH
    DERIVED_FROM
    INTERACTS_WITH
    BINDS_TO
    CATALYSES
    PARTICIPATES_IN
    LOCATED_IN
    HAS_FUNCTION
    ASSOCIATED_WITH
    AFFECTED_BY
    PREDICTS
    CLASSIFIES
    EMBEDS
    REPRESENTS
    TRAINED_ON
    VALIDATED_ON
    INFERRED_FROM
    SUPPORTED_BY
    CONTRADICTED_BY
    HAS_CONFIDENCE
    HAS_PROVENANCE
    MAPS_TO
    SEMANTICALLY_EQUIVALENT_TO
  }

  INPUTS {
    PROTEIN_SEQUENCES
    AMINO_ACID_SEQUENCES
    MULTIPLE_SEQUENCE_ALIGNMENTS
    STRUCTURAL_TEMPLATES
    EXPERIMENTAL_STRUCTURES
    PROTEIN_FAMILY_DATA
    DOMAIN_ANNOTATIONS
    FUNCTIONAL_ANNOTATIONS
    MUTATION_DATA
    INTERACTION_DATA
    LIGAND_DATA
    BIOLOGICAL_KNOWLEDGE_GRAPHS
    ONTOLOGY_TERMS
    EXPERIMENTAL_EVIDENCE
  }

  NORMALISE {
    STANDARDISE_PROTEIN_IDENTIFIERS
    NORMALISE_SEQUENCE_FORMATS
    NORMALISE_RESIDUE_NUMBERING
    RESOLVE_ISOFORMS
    MAP_DOMAINS_TO_CANONICAL_TERMS
    MAP_FUNCTIONAL_TERMS
    ALIGN_ONTOLOGY_TERMS
    REMOVE_DUPLICATES
    VALIDATE_METADATA
    PRESERVE_PROVENANCE
  }

  PROCESS PROTEIN_PREDICTION_PIPELINE {
    INPUT -> QUALITY_CONTROL
    QUALITY_CONTROL -> SEQUENCE_PREPROCESSING
    SEQUENCE_PREPROCESSING -> FEATURE_EXTRACTION
    FEATURE_EXTRACTION -> REPRESENTATION_LEARNING
    REPRESENTATION_LEARNING -> MODEL_INFERENCE
    MODEL_INFERENCE -> PREDICTION
    PREDICTION -> BIOLOGICAL_INTERPRETATION
    BIOLOGICAL_INTERPRETATION -> EVIDENCE_INTEGRATION
    EVIDENCE_INTEGRATION -> CONFIDENCE_ESTIMATION
    CONFIDENCE_ESTIMATION -> OUTPUT
  }

  PROCESS STRUCTURE_PREDICTION {
    PROTEIN_SEQUENCE
      -> SEQUENCE_REPRESENTATION

    SEQUENCE_REPRESENTATION
      -> RESIDUE_RELATIONSHIP_INFERENCE

    RESIDUE_RELATIONSHIP_INFERENCE
      -> SECONDARY_STRUCTURE_PREDICTION
      -> SPATIAL_CONSTRAINT_PREDICTION

    SPATIAL_CONSTRAINT_PREDICTION
      -> THREE_DIMENSIONAL_STRUCTURE_MODEL

    THREE_DIMENSIONAL_STRUCTURE_MODEL
      -> STRUCTURE_VALIDATION
      -> CONFIDENCE_SCORE
  }

  PROCESS FUNCTION_PREDICTION {
    PROTEIN_SEQUENCE
    + PROTEIN_STRUCTURE
    + DOMAIN
    + MOTIF
      -> FUNCTIONAL_FEATURE_MODEL

    FUNCTIONAL_FEATURE_MODEL
      -> MOLECULAR_FUNCTION_PREDICTION
      -> BIOLOGICAL_PROCESS_PREDICTION
      -> CELLULAR_COMPONENT_PREDICTION

    FUNCTION_PREDICTION
      -> EVIDENCE_INTEGRATION
      -> CONFIDENCE_SCORE
  }

  PROCESS INTERACTION_PREDICTION {
    PROTEIN_A + PROTEIN_B
      -> REPRESENTATION_PAIR

    REPRESENTATION_PAIR
      -> INTERACTION_MODEL

    INTERACTION_MODEL
      -> PROTEIN_PROTEIN_INTERACTION_PREDICTION
      -> INTERFACE_PREDICTION
      -> CONFIDENCE_SCORE
  }

  PROCESS PROTEIN_LIGAND_PREDICTION {
    PROTEIN_STRUCTURE + LIGAND
      -> BINDING_REPRESENTATION

    BINDING_REPRESENTATION
      -> BINDING_SITE_PREDICTION
      -> INTERACTION_PREDICTION
      -> AFFINITY_ESTIMATION

    AFFINITY_ESTIMATION
      -> CONFIDENCE_SCORE
  }

  PROCESS MUTATION_EFFECT_PREDICTION {
    PROTEIN + MUTATION
      -> ALTERED_SEQUENCE_OR_STRUCTURE

    ALTERED_SEQUENCE_OR_STRUCTURE
      -> STABILITY_PREDICTION
      -> FUNCTIONAL_EFFECT_PREDICTION
      -> INTERACTION_EFFECT_PREDICTION

    FUNCTIONAL_EFFECT_PREDICTION
      -> BIOLOGICAL_INTERPRETATION
      -> CONFIDENCE_SCORE
  }

  PROCESS KNOWLEDGE_GRAPH_INTEGRATION {
    PROTEIN + GENE + FUNCTION + STRUCTURE + INTERACTION + PATHWAY + EVIDENCE
      -> PROTEIN_KNOWLEDGE_GRAPH

    PROTEIN_KNOWLEDGE_GRAPH
      -> ENTITY_RESOLUTION
      -> LINK_PREDICTION
      -> RELATION_INFERENCE
      -> EXPLANATION_PATH
  }

  AI_TASKS {
    PROTEIN_SEQUENCE_EMBEDDING
    SECONDARY_STRUCTURE_PREDICTION
    TERTIARY_STRUCTURE_PREDICTION
    QUATERNARY_STRUCTURE_PREDICTION
    PROTEIN_FOLD_PREDICTION
    DOMAIN_PREDICTION
    MOTIF_PREDICTION
    BINDING_SITE_PREDICTION
    ACTIVE_SITE_PREDICTION
    PROTEIN_FUNCTION_PREDICTION
    PROTEIN_PROTEIN_INTERACTION_PREDICTION
    PROTEIN_LIGAND_INTERACTION_PREDICTION
    INTERFACE_PREDICTION
    STABILITY_PREDICTION
    MUTATION_EFFECT_PREDICTION
    PROTEIN_DESIGN_SUPPORT
    KNOWLEDGE_GRAPH_COMPLETION
  }

  RULES {
    IF ENTITY == PROTEIN
    THEN REQUIRE {
      CANONICAL_IDENTIFIER_OR_SEQUENCE
      BIOLOGICAL_CONTEXT
      PROVENANCE
    }

    IF AI_MODEL PREDICTS PROTEIN_STRUCTURE
    THEN LINK {
      PROTEIN_SEQUENCE
      PREDICTED_STRUCTURE
      CONFIDENCE_SCORE
      MODEL_PROVENANCE
    }

    IF MODEL_OUTPUT == PREDICTION
    THEN DO_NOT_ASSERT_AS_EXPERIMENTAL_OBSERVATION

    IF STRUCTURAL_SIMILARITY_EXISTS
    THEN DO_NOT_AUTOMATICALLY_ASSERT_IDENTICAL_FUNCTION

    IF SEQUENCE_SIMILARITY_EXISTS
    THEN DO_NOT_AUTOMATICALLY_ASSERT_IDENTICAL_STRUCTURE_OR_FUNCTION

    IF PREDICTED_BINDING_EXISTS
    THEN DO_NOT_AUTOMATICALLY_ASSERT_BIOLOGICAL_ACTIVITY

    IF MUTATION_EFFECT_IS_PREDICTED
    THEN PRESERVE_MODEL_UNCERTAINTY

    IF SEMANTIC_IDENTIFIERS_DIFFER
    AND BIOLOGICAL_REFERENT_IS_SAME
    THEN APPLY SEMANTIC_EQUIVALENCE
  }

  CONFIDENCE {
    CONFIDENCE_SCORE =
      WEIGHT(
        SEQUENCE_QUALITY,
        MODEL_VALIDATION,
        STRUCTURAL_SUPPORT,
        HOMOLOGY_SUPPORT,
        EXPERIMENTAL_SUPPORT,
        INDEPENDENT_REPLICATION,
        BIOLOGICAL_CONTEXT,
        PROVENANCE_COMPLETENESS
      )

    LEVELS {
      HIGH
      MODERATE
      LOW
      INSUFFICIENT
    }

    IF CONFLICTING_EVIDENCE
    THEN REDUCE CONFIDENCE_SCORE

    IF EXPERIMENTAL_STRUCTURE_SUPPORT
    THEN INCREASE CONFIDENCE_SCORE

    IF INDEPENDENT_FUNCTIONAL_VALIDATION
    THEN INCREASE CONFIDENCE_SCORE
  }

  SEMANTIC_EQUIVALENCE {
    EQUIVALENCE_KEY =
      NORMALISE(
        ENTITY_TYPE,
        CANONICAL_IDENTIFIER,
        SEQUENCE,
        ISOFORM,
        BIOLOGICAL_CONTEXT
      )

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

    ELSE PRESERVE_DISTINCT_ENTITIES
  }

  REASONING_CONSTRAINTS {
    PREDICTION != EXPERIMENTAL_OBSERVATION
    SEQUENCE_SIMILARITY != FUNCTIONAL_IDENTITY
    STRUCTURAL_SIMILARITY != FUNCTIONAL_IDENTITY
    BINDING_PREDICTION != BIOLOGICAL_ACTIVITY
    MODEL_CONFIDENCE != BIOLOGICAL_CERTAINTY
    PROTEIN_FUNCTION_IS_CONTEXT_DEPENDENT
    ISOFORMS_MUST_NOT_BE_CONFLATED
    RESIDUE_NUMBERING_MUST_BE_CONTEXTUALISED
    PROVENANCE_MUST_BE_PRESERVED
    UNCERTAINTY_MUST_NOT_BE_DISCARDED
  }

  VALIDATION {
    CHECK_ENTITY_TYPES
    CHECK_RELATION_DOMAINS_AND_RANGES
    CHECK_PROTEIN_IDENTIFIERS
    CHECK_SEQUENCE_INTEGRITY
    CHECK_ISOFORM_CONTEXT
    CHECK_STRUCTURE_REFERENCES
    CHECK_EVIDENCE_LINKS
    CHECK_PROVENANCE
    CHECK_CONFIDENCE
    CHECK_SEMANTIC_CONSISTENCY
    CHECK_DUPLICATE_ENTITIES
    CHECK_UNSUPPORTED_FUNCTIONAL_ASSERTIONS
  }

  OUTPUTS {
    NORMALISED_PROTEIN_ENTITIES
    SEQUENCE_REPRESENTATIONS
    PROTEIN_EMBEDDINGS
    PREDICTED_STRUCTURES
    STRUCTURAL_FEATURES
    FUNCTION_PREDICTIONS
    DOMAIN_PREDICTIONS
    MOTIF_PREDICTIONS
    BINDING_SITE_PREDICTIONS
    INTERACTION_PREDICTIONS
    MUTATION_EFFECT_PREDICTIONS
    KNOWLEDGE_GRAPH_EDGES
    CONFIDENCE_SCORES
    EVIDENCE_LINKS
    PROVENANCE_RECORDS
    EXPLANATION_PATHS
  }

  UPDATE_MODEL {
    ON NEW_PROTEIN_DATA {
      NORMALISE
      ENTITY_RESOLUTION
      RECOMPUTE_RELEVANT_FEATURES
      UPDATE_PREDICTIONS
      UPDATE_CONFIDENCE
      PRESERVE_VERSION_HISTORY
    }

    ON NEW_EXPERIMENTAL_EVIDENCE {
      LINK_EVIDENCE
      REASSESS_STRUCTURE_SUPPORT
      REASSESS_FUNCTION_SUPPORT
      REASSESS_INTERACTION_SUPPORT
      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_PROTEIN
      BY_SEQUENCE
      BY_STRUCTURE
      BY_FUNCTION
      BY_DOMAIN
      BY_INTERACTION
      BY_MUTATION
      BY_EVIDENCE
      BY_CONFIDENCE
    }

    REASON {
      APPLY_RELATIONSHIPS
      APPLY_RULES
      PRESERVE_UNCERTAINTY
      TRACE_EVIDENCE
      TRACE_PROVENANCE
      GENERATE_EXPLANATION_PATH
    }

    RESPONSE_POLICY {
      DISTINGUISH_PREDICTED_FROM_EXPERIMENTAL
      DISTINGUISH_SEQUENCE_FROM_STRUCTURE_EVIDENCE
      RETURN_CONFIDENCE_WHEN_AVAILABLE
      RETURN_PROVENANCE_WHEN_AVAILABLE
      EXPOSE_CONFLICTING_EVIDENCE
      AVOID_UNSUPPORTED_FUNCTIONAL_INFERENCE
    }
  }

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

END FLOW