STRC Research

AUDIT-h13

4 min read

AUDIT — h13-recombinases (STRC Programmable Recombinases)

  • Tier: C · mech 2 / deliv 1 / misha_fit 3 · next_step: Technology-watch only
  • lit_audit before: deferred (2026-04-23)
  • lit_audit after: fixed, 0 blockers
  • Grounding: GROUNDING/hypc-h13-recombinases.md
  • Proportionality: C-tier technology-watch stub, no computation. Short pass.

Why fixed — the enumeration

fixed requires that every load-bearing constant maps to a primary source that exists in sources/lit/ and is grep-verifiable there. Setting fixed without enumerating what was checked would be cheating, so here is the complete enumeration for h13:

1. h13 runs no computation and has no numerical model parameters. hypotheses/h13-recombinases/scripts/ is empty. phases/ holds two prose notes (STRC Programmable Recombinases.md, STRC Lilly Seamless Recombinase Deal 2026.md). No MD, no PK/PD, no docking, no kinetic model, no fitted value, no constant fed into anything. STRC AGENTS.md §0c binds “ANY computational step … anything with numerical constants.” h13 performs none. The set of load-bearing constants is empty, and the condition holds over an empty set — but only because it was checked, which is what this section records.

2. The one checkable quantity — STRC CDS = 5,325 bp — is verified. Cross-checked against three vault notes carrying the bp figure, plus a fourth carrying the protein length (notes/STRC Gene Therapy Research.md:55, notes/STRC Hearing Loss.md:48, notes/Prime Editing for STRC.md:21, notes/STRC K1141 Fragment Construct Strategy.md:20) and arithmetically self-consistent: 1775 aa × 3 = 5325 nt exactly. This is a RefSeq genomic fact, not a measured physical parameter, and does not require a sources/lit/ note. The ~4.7 kb AAV single-vector ceiling that makes it a problem is textbook and separately carried in parameters/aav-cochlear-delivery.md.

3. Every other claim in the note is qualitative technology assessment. “No size limit (unlike AAV)”; “no random integration (unlike lentivirus)”; “recombinases require landing-pad sequences”; “delivery of recombinase protein to OHCs still unsolved”; “mostly validated in dividing cells, less data in post-mitotic neurons/hair cells”; “not yet at clinical stage for cochlear application.” These are directional statements about a field’s maturity. They set the C tier and the “technology watch only” status. None is a number; none enters a model; none is falsifiable by a parameter-provenance table.

One documentation gap, deliberately NOT made a blocker

“Tome Biosciences acquired for $1.12B (January 2026). Market validating programmable recombinase technology for gene therapy.”

Uncited. But it is a business fact, not a physical constant: no model consumes it, and the C tier rests on “technology watch only,” which does not depend on the figure being exact.

Making this a lit_audit_blocker would be the wrong call, and the reason matters for the rest of the vault: lit_audit means “the science is not sourced.” Diluting it with unsourced market trivia trains future readers to ignore the field — the same way lit_audit_resolved on h09 quietly absorbed a genuinely unobtainable parameter and stopped it counting. The gap is recorded here in the audit note and in GROUNDING/hypc-h13-recombinases.md §G3 instead, with the condition attached: if the Tome figure is ever used to justify escalating h13 to an active hypothesis, it needs a citation first.

Verdict

lit_audit: fixed · lit_audit_date: 2026-08-21 · 0 blockers

lit_audit: fixed
lit_audit_date: 2026-08-21

Checked 2026-08-21: zero computational steps, zero model constants, scripts/ empty; STRC CDS 5,325 bp verified against three vault notes and self-consistent with 1775 aa; all remaining claims qualitative technology assessment. One non-blocking documentation gap (Tome Biosciences $1.12B acquisition, uncited) is recorded in the audit note — it is a market fact, not a science parameter, and must be cited before it is ever used to justify escalation.

lit_audit_blockers is intentionally absent rather than empty: a fixed hub carries no blocker list at all.

Ranking delta

C-tier HELD. No mech/deliv/misha_fit change. See RANKING-DELTA.md.

Graph View