§19 · the honest scoreboard
The honest scoreboard: what is verified, and what is firewalled out
The kit is green: 18 discriminant batteries PASS, the SOX9 anchor reproduces four NCBI γ/A4 verifies fully offline, and every battery is byte-identical on a second run (determinism). 27 absolute quantities are firewalled [O], each naming its obstacle. [V].
All 18 batteries pass; the RNA-machinery atlas (12 genes) and imprinted atlas (12 loci) reproduce SOX9-gated, the disease atlas (16 oncology + 9 neurodegeneration loci) reproduces cross-package with no tuning, and 42 wide-window A4 reads verify offline. The verified register is direction-only; 27 absolute magnitudes stay [O].
Eighteen batteries, all green
Every discriminant is measured out of the vendored substrate at measured γ, graded, and reproduced twice byte-identically. The gate re-checks the SOX9 anchor offline on every run; fail the anchor and the pipeline refuses to write.
PASS · R_rna_layer
PASS · RS_rna_species
PASS · A4_a4_layer
PASS · TC_two_channel
PASS · TG_env_to_germline
PASS · GE_germline_escapee
PASS · CH_coordinate_heritability
PASS · I_transgenerational_immunity
PASS · IM_immune_maturation
PASS · V_rna_vaccine
PASS · GT_gene_therapy
PASS · AM_a4_application_map
PASS · FM_rna_feasibility_map
PASS · DM_disease_feasibility_map
PASS · FV_feasibility_validation
PASS · PO_parent_of_origin
PASS · VK_rna_vaccine_kinetics
PASS · LV_lever_map
NCBI offline verify · RNA-machinery γ
(12 genes, anchor reproduced) · imprinted γ
(12 loci, anchor reproduced) · disease γ
(16+9 loci,
4-locus cross-package) · 42 wide-window A4 reads
· determinism: 2×sha256 identical
The one-sentence synthesis
SET (γ, genome) vs DRIVE (h, environment/RNA, applied at an A4 coordinate) is the same decomposition behind inheritance, immunity, vaccines and gene therapy: γ is the unwritable ruler, A4 is the writable structure that RNA targets and the environment inherits, and reversibility — a drive is reversible, a SET-edit is not — organises the therapeutic map.
The firewall, itemised
What the substrate does not fix, this volume does not claim. Each open item names the obstacle that keeps it open; promoting any of them requires new measured dynamics, never a louder claim. Every clinical application is handed to clinicians and regulators.
| id | item | obstacle |
|---|---|---|
| O-1 | absolute inherited phenotype magnitude (how much a parental exposure shifts a child's trait) | the substrate gives the inherited **sign** and the **ordering** of heritability by barrier; it has no scale mapping a drive amplitude to a trait unit. Firewalled. |
| O-2 | absolute payload → phenotype dose (copies of an RNA species → effect size) | RNA enters as an effective drive h; there is no substrate constant converting molecule count to drive units, so no absolute dose is claimed. (Track I-1 reads the *ordering* of species gains only.) |
| O-3 | absolute wild generation-count to loss (which exact generation an effect disappears in vivo) | TG5 fixes the **decay ordering** (RNA faster than a deep-barrier mark) and the *condition* (re-writing) for persistence; the in-vivo per-generation noise is not measured here, so the absolute generation index (F2/F3/…) is not asserted. |
| O-4 | absolute vaccine titre / quantitative correlate of protection | V-battery fixes the flip **direction**, the schedule **shape** (interior optimum), and transience/persistence; it has no scale from switch-state to antibody titre or protection probability. Handed to clinicians/regulators. |
| O-5 | absolute clinical interval (days between vaccine doses) | the optimum interval tracks the **measured protection half-life** in substrate time; mapping substrate time to calendar days needs an in-vivo rate that is not measured here. |
| O-6 | absolute gene-therapy dose, edit efficiency, and clinical efficacy/safety | GT fixes the lever **choice** (geometry threshold), the correction **sign**, and reversibility; absolute dose/efficiency/outcome are clinical and firewalled to clinicians/regulators. |
| O-7 | tolerable-drive cap as an absolute physiological bound (GT3) | the cap is **declared** as a structural parameter to demonstrate the decision rule, not measured; the real tolerability bound is clinical. The *rule* (drive-reachable → Lever B; beyond cap → Lever A) is [V]; the cap's absolute value is [O]. |
| O-8 | absolute reprogramming-noise amplitude D in vivo (TG2/TG3/I3) | D is fixed (not tuned per gene) to expose the **ordering** of survival by barrier; its true in-vivo value is not measured here. All TG/I conclusions are read as *orderings/signs*, which are invariant to the exact D within the regime. |
| O-9 | absolute per-species RNA gain (miRNA vs piRNA vs tsRNA vs m6A effect size) | RS reads the species SIGN and the production-stability ORDERING from machinery barrier; there is no substrate constant giving an absolute per-species gain. |
| O-10 | absolute PGC vs zygotic reprogramming noise amplitudes (GE2) | D_pgc / D_zyg are fixed at two distinct levels to expose the harsher-window-dominates ORDERING; their true in-vivo values are not measured here. The ordering/compound conclusions are invariant within the regime. |
| O-11 | absolute critical re-write rate w* in vivo (GE3) | GE3 fixes that a BOUNDARY exists (maintained above w*, transient below) from the measured per-generation loss; the calendar/molecular value of w* in the wild is not measured here. |
| O-12 | absolute loop-contact energy / loop occupancy at an A4 coordinate (A4-2) | the A4 read gives contact-COMPETENCE (same helical face) and the ORDERING of reachability; there is no substrate constant giving an absolute contact free energy or fractional loop occupancy. |
| O-13 | absolute A4 window scale / compartment boundary in vivo (A4-1,A4-4) | A4 is measured from ±15 kb windows; the true in-vivo compartment extent and 3D folding are not measured here. The orthogonality to γ and the coordinate-gating are invariant to the exact window within the regime. |
| O-14 | absolute parent-of-origin penetrance / contact stabilisation energy (CH1,CH2) | CH reads the coupling DIRECTION (contact helps at matched gamma) and the parent-resolved MAPPING; there is no substrate constant for an absolute contact stabilisation energy or parent-of-origin penetrance. |
| O-15 | absolute contact-assist energy delta and the calibrated boundary position (AM1,AM3) | AM reads that the drive-reachable boundary delta*(gamma) RISES with gamma and that contact-assist monotonically helps; the absolute assist energy and the calibrated boundary in physical units are not measured here. |
| O-16 | absolute selection pressure → affinity mapping in affinity maturation (IM1) | IM1 fixes the inverted-U **shape** (an interior-optimal selection pressure) and that maturation **emerges** from the germinal-centre loop; there is no substrate constant mapping the survivor fraction to a biological selection pressure, nor barrier depth to an antibody dissociation constant. The shape/emergence are [V]; the absolute pressure and affinity units are [O]. |
| O-17 | absolute innate vs adaptive memory lifetimes (IM2) | IM2 reads the MFPT **ratio** (two well-separated decay timescales, slow arm = deep barrier) in substrate time at a fixed noise; mapping substrate time to weeks (innate) or years (adaptive) needs an in-vivo escape rate that is not measured here. The separation/ordering are invariant within the rare-escape regime. |
| O-18 | absolute desensitisation dose/schedule and clinical tolerance outcome (IM3) | IM3 gives the memory↔tolerance **sign** and that **both** states are held basins after the drive clears (hysteresis); the number/size of tolerising exposures and the clinical desensitisation outcome are firewalled to clinicians and regulators. |
| O-19 | the self-contained yes/no of "does putting RNA in actually change the cell" (FM) | the answer rides on the RNA drive magnitude Δh (=dose/size), which is firewalled [O], on top of the assumed R19 dynamics — doubly undecidable self-contained. FM reads only the **map** (reachability=spinodal, ordering by measured γ, reversibility), all [V]; a "flipped" screen is the replay of the Δh assumption, not evidence. The yes/no is promoted only by **(B)**: scoring the predicted flipped-switch set+ordering against HELD-OUT measured pre/post expression of real siRNA/saRNA-treated cells, with no tuning of Δh or parameters to the target. |
| O-20 | whether a given RNA dose changes a given autism-gene switch in a real neuron (FM2 brain extension) | FM2 shows the **structure** transfers to MEASURED ASD promoters (each R19-bistable; flip-drive ordered by measured γ) — [V]; but the per-neuron yes/no rides on the firewalled Δh exactly as O-19, and is promoted only by **(B)** scored on held-out neuronal perturbation data (ASO/siRNA/saRNA series), no-tuning. Clinical RNA therapy is firewalled to clinicians and regulators. |
| O-21 | absolute corrective dose Δh per disease gene (how much drive restores a suppressor / silences an oncogene or toxic-GOF gene) (DM2/DM3) | DM fixes the corrective **sign** (mechanism-forced: LOF→`+`, GOF→`−`), the **ordering** of correction difficulty by measured γ, and that the corrective drive is **reachable** with a reversible Lever-B; it has **no** substrate constant mapping a drive amplitude to a clinical dose. The sign/ordering/lever-choice are [V]; the absolute dose is [O] and firewalled to clinicians and regulators. |
| O-22 | whether a given drive actually corrects a given patient's disease switch (the per-patient yes/no) (DM4) | exactly the O-19 circularity in the disease setting: a "corrected" screen exists at k≥1 for every switch by construction, so it is the replay of the firewalled Δh assumption on top of assumed R19 dynamics — doubly undecidable self-contained. DM reads only the **map** (reachability, sign law, ordering, reversibility), all [V]; the per-patient corrected-yes/no is promoted only by **(B)**: scoring the predicted corrected-switch set + ordering against HELD-OUT measured pre/post expression of real perturbed patient-derived cells, **no** tuning of Δh or parameters. All clinical translation firewalled to clinicians and regulators. **(B)-identifiability note (FV6):** the *ordering* path to this yes/no is non-identified (γ↔GC collinear, ρ=0.99), while the *sign* path (DM2) is **identified AND firewall-clean** (point-biserial(sign,GC)=0.0153) — so the precise obstacle is **bidirectional disease-correction DATA** (a `+` arm and a `−` arm), not the firewalled Δh. **(B) `−`-arm SCORED note (FV7, v0.17.0):** the `−` arm of that bidirectional test is now **scored** on held-out DepMap 24Q2 CRISPR-KO gene-effect — a knockout *is* the `−` operation and cancer-cell viability is its correction phenotype: GOF oncogenes (`corr_sign −`) are dependencies and LOF suppressors (`corr_sign +`) are not, **point-biserial(GOF, −gene_effect) = +0.4940, exact one-sided p = 0.0227** (n=16 onco, predicted direction, per-gene 12/16). It is **identified** (re-checked live: point-biserial(sign,GC) = −0.1222, GOF/LOF mean GC 0.532/0.543 balanced, GC-partialled +0.4939 unchanged) and **firewall-clean** (sign only). This earns the `−` arm a held-out `[V]` but **does not promote O-22**: only the `−` arm is in, so O-22's obstacle **narrows to the `+` RESTORE arm alone** (over-expression of a LOF suppressor / a toxic-GOF PD-gene knockdown with a disease/normal contrast). FV7 does **not** reintroduce Δh here — that magnitude is the separate item **O-21**. **(B) `+`-arm SCORED note — bidirectional test COMPLETE, obstacle CORRECTED (FV8, v0.18.0):** the `+` arm is now **scored** on held-out Horlbeck 2016 hCRISPRa-v2 K562 growth — a CRISPR-**activation** *is* the `+` operation and cancer-cell growth is its correction phenotype: LOF suppressors (`corr_sign +`) are growth-suppressive on activation and GOF oncogenes (`corr_sign −`) are not, **point-biserial(LOF, −growth_phenotype) = +0.4852, exact one-sided p = 0.0274** (n=16 onco, predicted direction, per-gene 10/16). It is **identified** (re-checked live: point-biserial(sign,GC) = +0.1222, LOF/GOF mean GC 0.5426/0.5320 balanced, GC-partialled +0.4731 unchanged) and **firewall-clean** (sign only). Together with FV7 the corrective sign-law is now **bidirectionally `[V]`** (both arms, direction-only), which **removes in full** the bidirectional-DATA obstacle. **O-22 STILL stays `[O]`**, and FV8 **corrects** FV7's "narrows to the `+` arm alone" framing: with both arms scored, O-22's residual obstacle is **no longer DATA — it is the firewall itself**. The sign-law is **class-level + direction-only** (which way to push each class); O-22 is a **per-patient, absolute** outcome needing the firewalled **per-patient magnitude/penetrance** (absolute drive size = **O-21**). A direction-only law cannot certify a per-patient yes/no, and promoting O-22 would leak direction-only `[V]` into per-patient absolute `[V]`. So O-22's obstacle **moves** from "missing `+`-arm data" to "the firewall (per-patient magnitude)" — not a contradiction of FV6 (the sign-law needed bidirectional data, not Δh; both arms scored with zero Δh). See FV7, FV8, and the "(B) validation" section below. |
| O-23 | absolute parent-of-origin penetrance (which parent's payload prevails in vivo, by how much) (PO1,PO2) | PO reads the context asymmetry (a SUSTAINED maternal drive crosses a held switch by DURATION while a TRANSIENT paternal burst of equal amplitude reverts), the inherited SIGN under opposing parents, and that the asymmetry is universal across the measured germline atlas — all [V]; there is no substrate constant mapping the sustained-vs-transient duration ratio to an in-vivo penetrance, so the absolute which-parent-wins probability is firewalled. |
| O-24 | absolute prime–boost interval in calendar days (VK1) | VK1 fixes the interval optimum **shape** (an interior optimum in germinal-centre rounds: too-short under-matures, too-long lets protection decay) from the measured maturation gain and the measured escape rate; mapping substrate GC-rounds to calendar days needs an in-vivo rate that is not measured here, so the absolute day count is [O] and clinical. (Same obstacle class as O-5, now for the maturation-derived interval.) |
| O-25 | absolute saRNA vs mRNA titre / durability magnitude (VK2) | VK2 reads only that the longer same-amplitude saRNA drive window crosses the protected (near-spinodal) basin **more reliably** than a shorter mRNA pulse — a reachability ORDERING — and notes honestly that post-flip durability is the **identical** barrier (the advantage is in reaching, not holding); there is no substrate constant giving an absolute titre or a durability in days, so both are firewalled. |
| O-26 | absolute per-modality lever efficiency (edit efficiency, knockdown depth, activation fold) (LV1,LV2) | LV1/LV2 classify each modality by **channel** (SET edit vs fixed-γ drive), **sign**, and **reversibility** — a falsifiable binning, all [V]; there is no substrate constant mapping a modality to an absolute edit efficiency, knockdown fraction, or activation fold, so the per-modality efficiency is [O] and firewalled to clinicians/regulators. |
| O-27 | absolute edit-free re-dosing schedule that maintains a correction in vivo (LV4) | LV4 fixes that a re-write BOUNDARY exists (a corrected state is MAINTAINED above a critical re-dose rate w* and FADES below it) from the measured per-cycle relaxation loss, with γ never touched; the calendar/molecular value of w* (how often a Lever-B drive must be re-dosed in the wild) is not measured here, so the absolute schedule is [O]. (Same obstacle class as O-11, now for therapeutic Lever-B re-dosing rather than germline re-writing.) |
FIREWALL · This volume reads WHICH switch, the SIGN of the drive, and the ORDERING / DECAY / BOUNDARY of effects; it never asserts an absolute magnitude. Every clinical application — vaccination, gene therapy, fertility care — is handed to clinicians and regulators. see the [O] ledger →