Data & episode context
Blood, fecal microbiota, urine biomarkers and metabolomics, genetics, wearables, authorized metadata and collection context.
Preserves: source value · unit · time · provenanceDeterministic platform architecture
A governed state layer - not another black-box score.
AXIS formalizes biological knowledge across multiple source families before statistics, recommendations or machine learning are applied. Each output remains tied to source provenance, a governed module, version and evidence path.
01 / Core hierarchy
From multiple biological sources to a traceable state.
The hierarchy is reusable across products. Every source retains provenance, and every module defines its own within-domain patterns, cross-patterns, axes, claims and validation state.
Canonical parameters
Stable identifiers, comparable units, provenance, transformation and module-specific System Insight Range (SIR).
Adds: canonical identity · transformation · module SIRDomain patterns & cross-patterns
Explicit deterministic activations within one biological domain or across two or more source families.
Evaluates: activation · coverage · confidence · coherenceBlocks
Related patterns are organized with confidence logic, redundancy control and conflict resolution.
Handles: contribution · redundancy · directional conflictAxes
Higher-order biological-state dimensions specific to each governed module.
Produces: governed higher-order state dimensionsDBSI & trajectory
DBSI means Deterministic Biological State Index: an explainable 0–100 episode synthesis, with confidence, coverage and longitudinal comparison preserved.
Retains: index · axes · confidence · trajectory · evidence path
A cross-pattern is an explicit, versioned relationship whose evidence comes from two or more biological domains.
Blood, fecal microbiota, urine biomarkers and metabolomics, genetics, wearables and phenotypic and contextual data remain source-identifiable while contributing to convergence, discordance or interaction logic.
02 / Core invariants
What AXIS must preserve as it scales.
Commercial modules can change. The Core principles cannot change silently.
- Same input + same version = same output.
- Every output retains its calculation path and configuration version.
- Parameter identity is canonical; interpretation is module-specific.
- Confidence, coverage, coherence and conflict logic remain visible.
- No machine-learning layer may silently rewrite deterministic rules or weights.
System Insight Range (SIR)
Context before interpretation.
SIR defines the module-specific interval used to interpret a canonical parameter. It is versioned with the module and remains distinct from the source laboratory reference interval.
Canonical identityOne parameter definition across the platform.
Module interpretationDifferent SIR and descriptions where scientifically justified.
03 / Confidence & conflicts
Explainability is part of the computation - not a paragraph added afterwards.
AXIS separates signal strength from confidence and preserves coverage, key-marker sufficiency, coherence and competing biological directions.
✓
Coverage
How much of the required module evidence is actually present.
Explore +✓
Key-marker sufficiency
Whether the most decisive evidence is available.
Explore +✓
Coherence
Whether contributing signals support a consistent biological direction.
Explore +✓
Conflict handling
How redundancy, opposition and dominant evidence affect aggregation.
Explore +01 / COVERAGE
Coverage
Coverage describes how much of the evidence expected by a declared module is actually available for execution.
Illustrative example
If an expected data family or required group of measurements is missing, coverage decreases even when the available signals point in the same direction.
Why it mattersA strong activation with incomplete coverage must remain visibly different from the same activation supported by a complete evidence set.
04 / Separated domains
One governed Core - Distinct intended-use pathways.
Shared infrastructure does not imply shared claims. Medical, research and wellness configurations remain independently versioned and governed.
Evidence and regulatory execution
Medical modules require their own validation, quality system, labeling and jurisdiction-specific regulatory pathway.
Provider-supervised non-diagnostic workflows
Wellness configurations can organize and contextualize biological information without diagnosing disease or replacing professional judgment.
Research
Exploratory configurations with explicit lifecycle states
Research logic can explore candidate relationships and cohort states without inheriting clinical validity or commercial claims.
05 / AI relationship
ML can learn from AXIS - ML does not define AXIS.
AXIS exposes stable, versioned state vectors to downstream analytics. Those systems can learn from the output, but they cannot modify the parameters, rules, weights or inference logic inside the deterministic Core.
- Use AXIS state vectors as interpretable downstream features.
- Keep experimental models outside the deterministic Core.
- Declare training data, model version, purpose and validation status.
- Preserve the original state and evidence path for every derived result.
Layered architecture · controlled interfaces
Analytics, recommendations & ML
Models may consume governed AXIS state vectors for prediction, stratification or decision support.
Separate governance: training data · purpose · model version · validationProduct interpretation
Descriptions, reports, timelines and module-specific views translate the state without redefining it.
Consumes Core outputs · never alters inferenceDeterministic Core
Canonical parameters and explicit rules execute the path from patterns to blocks, axes and state.
Versioned · reproducible · traceable to source evidenceArchitecture review
Evaluate AXIS at the level that matters: rules, evidence and reproducibility.
Qualified partners can review a bounded technical demonstration and discuss the module contract required for their use case.