GROUNDING — h14 protein replacement (proportionate pass, C-tier)

G1 — the hub’s kill contradicts the note’s own analysis

Hub frontmatter, hypotheses/h14-protein-replacement/index.md:

next_step: No delivery route

Body of phases/STRC Protein Replacement Therapy.md:64:

Diffusion modeling — Stokes-Einstein: at 115 kDa, D ≈ 2-4 µm²/s in aqueous. Perilymph ~5 mm long. Time to diffuse to all OHC rows: hours. Feasible.

and line 50:

ERT scale: Modern CHO/HEK cell expression systems routinely produce large glycoproteins at therapeutic scale. Mini-STRC (115 kDa) is smaller than several approved ERT proteins.”

Evidential weight: INTERNAL INCONSISTENCY, recorded as-is. The hypothesis note concludes intracochlear distribution is feasible in hours; the hub kills the hypothesis for having no delivery route. Both may be reconcilable — “no delivery route” plausibly means getting protein across the round-window membrane into perilymph in the first place, which is a different step from diffusing along perilymph once there — but that reconciliation is nowhere written down. As the vault stands, a reader cannot tell whether h14 was killed on evidence or on an unstated assumption. Auditing what a kill rests on is exactly this route’s job for a C/D-tier hypothesis, so this is the finding.

G2 — the RWM step is already sourced elsewhere and was never applied here

sources/lit/2001-salt-ma-quantification-rwm-permeability.md (type: lit) was retrieved, MinerU-parsed and closed to ✅ during the h09 audit. It carries, at line 44:

| RWM permeability (TMPA) | 1.9 × 10⁻⁸ | cm/s | Guinea pig; TMPA ion MW ~166 Da |

and its own MW-scaling derivation at lines 61–62:

“Using D ∝ r⁻¹ ∝ MW^−1/3 (spherical molecule): P(14 kDa) = 1.9×10⁻⁸ × (166/14000)^(1/3) ≈ 4.9×10⁻⁹ cm/s” “Using D ∝ MW^−1/2 (polymer/chain scaling): P(14 kDa) = 1.9×10⁻⁸ × (166/14000)^(1/2) ≈ 2.1×10⁻⁹ cm/s”

Evidential weight: STATES — for h09’s parameter; VARIANT for h14’s, requiring one further extrapolation. Extending the same MW^−1/3 scaling from 14 kDa out to mini-STRC at 115 kDa gives P ≈ 1.9×10⁻⁸ × (166/115000)^(1/3) ≈ 2.2 × 10⁻⁹ cm/s, roughly an order of magnitude below the TMPA ion. This extrapolation is presented here as an illustration of how the kill could be quantified, not as a sourced result: Salt & Ma measured a 166 Da ion, and pushing a Stokes–Einstein scaling across nearly three decades of molecular weight to a heavily glycosylated 115 kDa glycoprotein is well outside what that paper supports. It is exactly the kind of extrapolation §0c exists to stop from hardening into a “value.”

The point stands regardless of the arithmetic: h14’s kill is quantifiable from a paper the vault already owns, and nobody has done it. That is a cheap, concrete next step, not a literature gap.

G3 — the uncited diffusion coefficient

“at 115 kDa, D ≈ 2-4 µm²/s in aqueous”

No citation. It is an order-of-magnitude-plausible Stokes–Einstein estimate for a protein of that size, and the note says “Stokes-Einstein” so the method is at least declared. But no sources/lit/ file backs it and no radius is given for the calculation.

Evidential weight: UNSOURCED ESTIMATE. It is load-bearing for the “feasible” half of G1’s contradiction, so it cannot be waved through.

G4 — STRC mass figures

“STRC is large (~220 kDa full-length, ~115 kDa for mini-STRC 700-1775)”

Consistent with 1775 aa full-length (see GROUNDING/hypc-h13-recombinases.md §G2) and with mini-STRC spanning residues 700–1775 (1076 aa). Average residue mass ~110 Da gives ~195 kDa and ~118 kDa respectively; the stated 220 kDa reflects STRC’s heavy N-glycosylation, which the note itself flags at line 67. Internally coherent; a derived quantity, not a measurement. Not a blocker.