Add genetics information substrate boundary

0-Core-Formalism/otom/docs/genetics_information_substrate_boundary.md

@@ -0,0 +1,163 @@
+---
+project: OTOM
+domain: axis-04-formalization
+type: BoundaryRecord
+settlement: FORMING
+authority: registry
+route_signature: otom/axis-04-formalization/boundaryrecord/genetics-information-substrate/v0
+claim_state: BEAUTIFUL_PROVISIONAL
+source_audit: uploaded-codemap-2026-05-05
+---
+
+# Genetics Information Substrate Boundary
+
+## Purpose
+
+This boundary record separates genuinely plumbed genetics-information models from
+registry-only, ghost, aspirational, and misfiled entries.
+
+The immediate goal is claim-state hygiene:
+
+```text
+Do not claim the genetics layer is fully implemented when only part of it has
+compiled Lean, running data code, or executable evidence receipts.
+```
+
+## Canonical boundary statement
+
+```text
+The Research Stack currently has a partial but real genetics-information
+substrate: compiled codon/peptide/compression formalizations, one real-data
+population-genetics script, and a larger registry of unplumbed or ghost model
+entries. Claims about genetics information capacity must cite the plumbed
+substrate or remain BEAUTIFUL_PROVISIONAL.
+```
+
+## Classification buckets
+
+| Bucket | Meaning | Promotion rule |
+|---|---|---|
+| `CANONICAL_PLUMBED` | compiled Lean module or running real-data script | may support bounded claims |
+| `REGISTRY_ONLY` | equation/model exists in registry but has no executable path | may support roadmap claims only |
+| `GHOST_ENTRY` | marked complete/available but no corresponding implementation found | must be demoted or implemented |
+| `MISFILED` | present near genetics but not genetics-relevant | relocate or mark out-of-scope |
+| `ASPIRATIONAL_MIXTURE_PRIMITIVE` | taxonomy concept without proof/data path | BEAUTIFUL_PROVISIONAL only |
+
+## Plumbed genetics substrate
+
+### Lean / formalized substrate
+
+The audit identifies the following as the strongest plumbed genetics-connected
+information models:
+
+| Module | Information-holding mechanism | Genetics tie | Status |
+|---|---|---|---|
+| `GeneticCode.lean` | NCBI Table 1 codon-to-amino-acid map; 64 codon states plus outputs | canonical translation and codon degeneracy | `CANONICAL_PLUMBED` |
+| `CodonOTOM.lean` | codon efficiency functional, mutation delta, selection condition | codon usage bias and mutation-selection balance | `CANONICAL_PLUMBED` |
+| `PeptideMoE.lean` | peptide state, Ramachandran coordinates, internal energy, entropy | conformational search as information geometry | `CANONICAL_PLUMBED` |
+| `GenomicCompression.lean` | compression windows, phiGenomic, Q16_16 arithmetic | formal sequence-compression substrate | `CANONICAL_PLUMBED` |
+| `SyntheticGeneticCoding.lean` | coding and biological parameter projections | normalized biological parameter projection | `CANONICAL_PLUMBED` |
+| `GeneticGroundUp.lean` | nucleotide-state encoding, expression probability, binding energy, fold angle | synthetic biology / nucleotide-state model | `CANONICAL_PLUMBED` |
+| `HachimojiPipeline.lean` | 8-symbol expanded genetic alphabet analog | expanded DNA/RNA coding regime | `CANONICAL_PLUMBED` |
+
+### Running data substrate
+
+| File | Information-holding mechanism | Genetics tie | Status |
+|---|---|---|---|
+| `Allelica.py` | Hardy-Weinberg genotype-frequency model over real gnomAD data | p² + 2pq + q² = 1 applied to real allele/genotype data | `CANONICAL_PLUMBED` |
+
+## Registry-only model surface
+
+These models may be scientifically relevant but should not be promoted as
+implemented until a Lean module, Python script, test fixture, or receipt path
+exists.
+
+| Model family | Examples | Status |
+|---|---|---|
+| Population genetics registry | Wright-Fisher drift, Fisher fundamental theorem, Price equation | `REGISTRY_ONLY` |
+| Sequence information registry | Shannon entropy, per-position entropy, Jensen-Shannon divergence, Jukes-Cantor distance, ANI/AAI, mutation rate, novel mutation count | `REGISTRY_ONLY` unless connected to compiled module |
+| Thermodynamic/folding registry | RNA folding delta-G, Gibbs free energy, fitness-entropy tradeoff, Waddington potential | `REGISTRY_ONLY` |
+| Cellular/regulatory registry | Central Dogma ODE, Gierer-Meinhardt pattern formation | `REGISTRY_ONLY` |
+
+## Ghost entries requiring demotion or implementation
+
+The audit reports these as present in the registry but not backed by a found Lean
+module, Python implementation, or data pipeline:
+
+| Ghost entry | Why it matters | Recommended action |
+|---|---|---|
+| Quasispecies equation | mutation-selection balance, error threshold, viral evolution | implement Lean skeleton + simulator |
+| Replicator equation | evolutionary game dynamics | implement only if tied to population-genetics or ecology use case |
+| Neutral theory / Kimura | drift-mutation balance | implement after drift/selection metrics |
+| Masked language modeling loss for DNA | sequence-prediction information model | defer until corpus/data pipeline exists |
+| pTM / TM-score | structural biology information metric | defer until protein-structure pipeline exists |
+
+## Misfiled entry
+
+| Entry | Reason | Action |
+|---|---|---|
+| `parametric-learn` | Three.js/TensorFlow.js parametric surface fitting demo; no biological sequence, population, or genetics-information role found | mark `MISFILED` or relocate outside genetics |
+
+## Missing canonical gaps
+
+| Missing area | Why it matters | Suggested implementation route |
+|---|---|---|
+| Selection detection | identifies signatures of selection in sequence/population data | implement Tajima's D and FST first |
+| Phylogenetic tree inference | reconstructs evolutionary history as tree-structured information | use Jukes-Cantor distance, then neighbor-joining |
+| Population structure | ancestry/latent-variable information | PCA/admixture after real genotype fixtures exist |
+| Epigenetic HMM / chromatin states | regulatory information layer | HMM implementation plus receipt schema |
+| Splicing / isoform quantification | transcript-level information | defer until RNA-seq or transcript fixtures exist |
+| Variant calling | raw read to variant state inference | hard; requires aligner/error model |
+| Comparative genomics | synteny/CNE/regulatory homology | hard; defer |
+
+## Mass Number interpretation for genetics
+
+Genetics is a communication regime, not a human-language regime.
+
+```text
+DNA sequence
+  -> transcriptional decoder
+  -> RNA folding / splicing decoder
+  -> codon translation decoder
+  -> peptide folding decoder
+  -> cellular phenotype decoder
+  -> population-selection decoder
+```
+
+Recommended Mass Number classes:
+
+```text
+MN-BIO-GEN   Genetic sequence mass
+MN-BIO-COD   Codon / translation mass
+MN-BIO-FOLD  Folding / conformation mass
+MN-BIO-POP   Population-genetic mass
+MN-BIO-REG   Regulatory / epigenetic mass
+MN-BIO-EVO   Evolutionary trajectory mass
+```
+
+## Promotion policy
+
+A genetics model may be promoted beyond `BEAUTIFUL_PROVISIONAL` only when it has
+at least one of:
+
+1. compiled Lean theorem or executable definition;
+2. running Python/Rust pipeline with test fixtures;
+3. real-data receipt with source provenance;
+4. adversarial or benchmark receipt;
+5. explicit connection to MassNumber admissibility or SemanticMass carrier scoring.
+
+## Immediate promotion candidate
+
+Implement selection metrics first:
+
+```text
+Tajima's D + FST
+```
+
+Rationale:
+
+- low implementation difficulty;
+- high genetics-information value;
+- direct path from allele/genotype state to population selection signal;
+- good MassNumber gate candidate for deciding whether selection evidence is
+  strong enough to promote a registry model.