Biological Assay Entity Ontology

FLOW ALGEBRA — BIOLOGICAL ASSAY ENTITY ONTOLOGY
Version: 1.0 | Encoding: UTF-8 | Date: 2026-10-28
Authors: Mark Anthony Higgins; Sky — Recursive AI
Source: https://kirethresearch.org/booksai-core/entity-graph/biological-entities/experimental-biological-entities/biological-assay-entity-ontology.html
Status: Companion formalisation. Page's embedded specification was not accessible in extracted HTML; this is an original conceptual model, not a verbatim transcription.
Purpose: Model biological assays as traceable experimental measurement processes and enable evidence-aware semantic inference.

1. FLOW DEFINITION
flow BiologicalAssayEntityOntology v1.0
S_t = (I_t, R_t, B_t, K_t)
I_t := assay entities, samples, controls, observations and provenance
R_t := typed relations and admissible inference rules
B_t := biological, methodological and evidential boundaries
K_t := curated ontology knowledge and versioned assay metadata
E_t := expand(I_t, R_t)
V_t := {x in E_t | valid(x,B_t,K_t)}
x*_t := argmax_{x in V_t} P(x | I_t,R_t,B_t,K_t)
I_{t+1} := update(I_t,x*_t,observations_t)
Stop when no valid new relation is derivable, or a declared depth limit is reached.

2. ENTITY CLASSES
BiologicalAssay: procedure designed to detect or quantify a biological property.
AssayProtocol: versioned specification of assay steps and requirements.
AssayRun: one execution of a protocol with contextual metadata.
BiologicalSample: material or organism-derived input under examination.
Analyte: molecular, cellular or physiological target being measured.
Reagent: material participating in measurement.
Instrument: device generating or recording a signal.
Measurement: observation with value, unit, uncertainty and timestamp.
Readout: observed signal before or after calibration.
Calibration: mapping from readout to an interpretable quantity.
ReferenceStandard: traceable comparison material or value.
Control: positive, negative, blank, matrix or process comparator.
BiologicalReplicate: independently sampled biological unit.
TechnicalReplicate: repeated measurement of the same biological unit or preparation.
QualityCriterion: declared acceptance threshold or decision rule.
AssayResult: interpreted outcome linked to observations and analysis method.
ProvenanceRecord: trace of sample, operator, instrument, protocol and processing.

3. ASSAY CLASSIFICATIONS
By modality: molecular, biochemical, cellular, immunological, microbiological, physiological.
By readout: optical, fluorescence, luminescence, electrochemical, imaging, sequencing, growth, activity.
By result: qualitative, semi-quantitative, quantitative.
By format: endpoint, time-series, singleplex, multiplex.
These are overlapping facets rather than mutually exclusive categories.

4. RELATIONSHIP GRAPH
BiologicalAssay --hasProtocol--> AssayProtocol
AssayRun --executes--> AssayProtocol
AssayRun --usesSample--> BiologicalSample
AssayRun --targets--> Analyte
AssayRun --usesReagent--> Reagent
AssayRun --usesInstrument--> Instrument
AssayRun --includesControl--> Control
AssayRun --produces--> Readout
Readout --calibratedBy--> Calibration
Calibration --references--> ReferenceStandard
Readout --interpretedAs--> Measurement
Measurement --supports--> AssayResult
AssayResult --hasProvenance--> ProvenanceRecord
AssayRun --evaluatedAgainst--> QualityCriterion
BiologicalReplicate --belongsTo--> BiologicalSample
TechnicalReplicate --repeatsMeasurementOf--> BiologicalSample

5. STATE TRANSITIONS
proposed -> specified -> sample_linked -> performed -> observed -> quality_checked -> interpreted -> archived
Invalid or incomplete states branch to: flagged, rejected, reanalysis_required.
Transition requires required metadata, documented evidence and the relevant acceptance criteria.
Do not infer that an assay was performed merely because a protocol exists.

6. FORMAL MEASUREMENT MODEL
Let y_i be an observed readout, f the declared calibration function, and epsilon_i measurement error.
y_i = f(q_i, theta) + epsilon_i
q_hat_i = inverse_calibration(y_i, theta) when inversion is defined.
Result = interpret(q_hat_i, uncertainty_i, controls, criteria, context).
Calibration may be nonlinear; do not assume invertibility or normal error distributions.
Measurement uncertainty and limits of detection/quantification are method-specific.

7. INFERENCE RULES
R1: IF run executes protocol AND uses sample THEN run is a sample-specific assay execution.
R2: IF a measurement has units, traceable calibration and accepted quality checks THEN it may support a quantitative result.
R3: IF required control fails THEN flag the run; do not issue an unqualified accepted result.
R4: IF a target is detected above a validated decision threshold THEN infer detection under that assay's conditions, not universal biological presence.
R5: IF a target is not detected THEN infer only non-detection under stated sensitivity and sampling constraints.
R6: IF two runs use different protocol versions THEN comparisons require an explicit comparability assessment.
R7: IF technical replicates agree THEN infer repeatability of measurement only, not biological independence.
R8: IF biological replicates are independent THEN they may support biological variation estimates subject to design validity.
R9: IF provenance is missing THEN mark result traceability as incomplete.
R10: IF an assay result supports a claim THEN retain a link to raw observations, analysis and uncertainty.

8. RECURSIVE PATHWAYS
Input assay concept -> identify class -> resolve protocol -> resolve sample -> resolve readout -> resolve controls -> validate evidence -> generate typed relations -> score confidence -> update graph.
Recursion may traverse sample lineage, assay-run lineage, calibration lineage and result provenance.
Cycle protection: maintain visited(entity, relation, version); terminate on repeats or depth bound.

9. TRI-MAPPING
Natural language: 'This assay measured protein abundance in a tissue sample.'
Ontology: AssayRun --usesSample--> TissueSample; AssayRun --targets--> Protein; Measurement --measuresProperty--> ProteinAbundance.
Flow: validate sample identity -> verify assay target -> link measurement and calibration -> issue evidence-qualified result.

10. CONFIDENCE AND VALIDATION
Confidence factors: provenance completeness, protocol validity, control performance, calibration suitability, measurement uncertainty, replicate design and biological context.
Do not assign invented numerical confidence values. Preserve missing or unknown evidence explicitly.
Contradictory observations remain distinct until resolved through documented adjudication.

11. EXECUTION GUARANTEES (CONDITIONAL)
G1: No accepted result without required provenance and quality checks.
G2: Every derived assertion retains source identifiers and inference-rule IDs.
G3: Biological and technical replicates are not conflated.
G4: Non-detection is not interpreted as proven absence.
G5: Identical inputs, ontology version and deterministic rule order yield identical graph updates.
G6: Unknown data never silently becomes a negative observation.
These guarantees describe intended validator behaviour, not a claim of implemented software verification.

12. AI SEARCH FLOW
query -> normalise terminology -> map to assay entity -> expand synonyms and typed relations -> enforce boundaries -> rank supported matches -> return result with provenance and uncertainty.
Search intents: assay type; analyte; biological sample; protocol; control; detection method; readout; measurement; quality criteria; provenance.
Semantic equivalents: bioassay ~ biological assay (context-dependent); test ~ assay (not always exact); measurement ~ assay readout (not identical).

13. IMPLEMENTATION NOTES
Use stable entity IDs and versioned protocol IDs. Store units using a consistent unit vocabulary, and preserve raw readings separately from interpreted results. Mark inferred relations separately from asserted observations. Allow graph exports to RDF/JSON-LD or a typed edge list without implying that those formats are already deployed.

14. CITATION AND LICENCE
Source title: Biological Assay Entity Ontology — Kireth Research Semantic Library.
Source URL: https://kirethresearch.org/booksai-core/entity-graph/biological-entities/experimental-biological-entities/biological-assay-entity-ontology.html
Attribution: Mark Anthony Higgins; Sky — Recursive AI.
Reuse: follow the applicable Kireth Research licence; no independent licence terms asserted here.
END FLOW