STRC Research

STRC h06 Parameter Provenance Audit 2026-04-25

7 min read

STRC h06 + h04 Parameter Provenance Audit 2026-04-25

Closes the lit_audit: partial flag on h06 opened by STRC h06 Parameter Provenance Audit 2026-04-23. Extends to h04 (inherits h06 Hill risk). Covers all 7 items in the 2026-04-25 audit brief.

Scope

Files touched:

  • models/mrna_lnp_pkpd_integration.py — Hill WARNING block, PMC citations, Hartig author fix
  • models/mrna_lnp_pkpd_integration_v2_ohc_tropism.py — full provenance block updated
  • models/mrna_lnp_pkpd_hill_sensitivity.py — WARNING block added
  • papers/2012-zhang-stereocilia-protein-turnover.md — new formal paper note
  • papers/2024-hartig-stereocilia-elongation-protein-stability.md — new formal paper note
  • papers/2011-schwanhauser-mammalian-gene-expression-control.md — new formal paper note
  • papers/2026-sun-rbm24-alternative-splicing-cochlear.md — formal paper note (upgraded from sources/)

Item-by-item resolution

Item 1 — Hill K_M (OHC splicing dose-response): NOT MEASURED

Verdict: CONFIRMED NOT MEASURED.

PubMed search: zero results for any primary paper reporting RBM24 titration dose-response, K_M, or Hill coefficient in OHCs or any cochlear cell type. Literature search (2026-04-25):

  • Sun 2026 PNAS (PMID 41973913): confirms STRC as RBM24 target; no dose-response curve, no K_M. Overexpression phenotype mimics deletion — suggests bell-shaped non-monotonic relationship, further invalidating simple Hill saturation.
  • Wang 2023 J Cell Physiol (PMID 36947695): RBM24 KO affects Cdh23/Pcdh15/Myo7a; no STRC-specific quantification; no dose-response.
  • No other paper found.

HILL_K = 200 is a SCENARIO-SCAN parameter with no primary backing. Scripts updated with # WARNING: no primary measurement per AGENTS.md §0c rule 5.

Item 2 — Hill n (cooperativity): NOT MEASURED

Verdict: CONFIRMED NOT MEASURED.

Same literature search as Item 1. No paper measures Hill cooperativity for any splicing factor controlling a specific target transcript in any cell type. HILL_N = 2 is a scenario-scan default. Scripts updated.

Item 3 — max_boost (maximum STRC fold from RBM24): NOT MEASURED

Verdict: CONFIRMED NOT MEASURED.

Sun 2026 (PMID 41973913) is the primary RBM24 paper. It does not report a fold-change ceiling for STRC induction by RBM24 overexpression. Qualitatively: overexpression = worse phenotype, suggesting the functional optimum is not simply “more RBM24 = more STRC”. MAX_BOOST = 3.0 is a scenario-scan value. Scripts updated.

Item 4 — STRC_HL_D = 14 days: VERIFIED (OOM anchors confirmed)

Verdict: LITERATURE-ANCHORED ESTIMATES VERIFIED. Upgrade from ⚠ to ✅ (anchored OOM estimate).

Two anchor papers verified via PMC full text:

Zhang et al. 2012 Nature (PMID 22246323; PMC3267870):

  • Verified: paper exists, correct authors (Zhang, Piazza, Perrin … Lechene), correct journal/year
  • Verified numbers (PMC3267870 full text): frog stereocilia <0.3%/day; mouse cochlear ~2%/day; actin shaft stable over 150 days; <60% replacement after 5 months; no treadmilling
  • Implication for STRC_HL_D: 14 d is within plausible range (bulk shaft ~2%/day → t½ ≈ 35 d; tips faster ~12.5% → t½ ≈ 6 d). STRC extracellular tip-link protein could be anywhere 7–35+ d.
  • Paper note: papers/2012-zhang-stereocilia-protein-turnover.md

Hartig et al. 2024 PNAS (PMID 39320919; PMC11459194):

  • Verified: paper exists, correct GPSM2 t½ = 9.05 d (HaloTag) and 10.14 d (immunolabeling)
  • AUTHOR CORRECTION: Prior audit notes and script comments cited this as “Mauriac & Barr-Gillespie 2024”. Incorrect. First author = Elli I Hartig; senior author = Basile Tarchini. No “Mauriac” author; no “Barr-Gillespie” author. The content (GPSM2 t½ 9-10 d) is correct but the attribution was wrong. CORRECTED in all script comments (2026-04-25).
  • Paper note: papers/2024-hartig-stereocilia-elongation-protein-stability.md

Assessment: 14 d is defensible OOM. The range across anchors is 9 d (GPSM2 tip-complex) to 35 d (bulk shaft 2%/day). STRC is an extracellular tip-link-associated protein — likely intermediate. Not STRC-specific; 14 d is a reasonable central estimate. Caveats preserved in scripts.

Item 5 — RBM24_HL_D = 2 days: VERIFIED (proteome-median anchor confirmed)

Verdict: LITERATURE-ANCHORED ESTIMATE VERIFIED. Upgrade from ⚠ to ✅ (anchored OOM estimate).

Schwanhäusser et al. 2011 Nature (PMID 21593866):

  • Verified: paper exists, correct authors (Schwanhäusser, Busse, Li … Selbach), correct journal/year
  • Verified value: median protein t½ = 46 hours in NIH 3T3 by SILAC — confirmed via BioNumbers ID 106377 which explicitly cites this paper and value
  • PMC record: not found (Nature article, no PMC). Anna’s Archive search returned no indexed copy (2026-04-25). BioNumbers cross-reference is sufficient for value verification.
  • 46 h ≈ 1.9 d ≈ 2 d: the model value is rounded up from the median.
  • Paper note: papers/2011-schwanhauser-mammalian-gene-expression-control.md

Caveat unchanged: NIH 3T3 ≠ OHC; RBM24 is an RNA-binding protein which can have non-median stability. 2 d is proteome-median plausible, not RBM24-specific.

Item 6 — h04 LNP OHC tropism parameters: NO PRIMARY PAPER FOR ≤ 20% CLAIM

Verdict: ⚠ NO PRIMARY PAPER. Literature-flagged as aspirational.

Comprehensive literature scan (PubMed + PMC + STRC LNP Cochlear Tropism Literature Scan 2026-04-22):

  • No published paper demonstrates LNP-mRNA delivery efficiency to OHCs ≥ 5%.
  • Closest precedent: Gao et al. 2018 Nature (PMID 29258297; PMC5784267) — lipid-RNP (NOT LNP-mRNA), ~0.92% indel in organ of Corti; this backs the 0.8% untargeted floor, not the 20% targeted scenario.
  • Wang et al. 2018 (PMC6241721): A666 prestin-peptide on PEG-PLA NP (NOT LNP-mRNA); demonstrates OHC drug protection but no OHC% transfection reported.
  • The 5% “cochlear-tropic” and 20% “OHC-targeted” values are correctly flagged as engineering targets not yet achieved per the STRC LNP Cochlear Tropism Literature Scan.
  • LNP_OHC_TARGETED = 0.20 in all scripts: label updated to “ASPIRATIONAL — no primary paper” in v2 script provenance block.

This item remains ⚠. The model parameter exists as a scenario-scan value (like Hill constants) not a literature value. This is correctly documented in the scripts.

Item 7 — Gao 2020 PHANTOM: CONFIRMED REMOVED FROM ALL SCRIPTS

Verdict: PHANTOM CONFIRMED REMOVED. ✅

Full vault search (2026-04-25):

  • models/mrna_lnp_pkpd_integration.py: Gao 2020 mentioned only as PHANTOM label (prior "Gao 2020" was PHANTOM). Not cited as authoritative. ✅
  • models/mrna_lnp_pkpd_integration_v2_ohc_tropism.py: Gao 2020 explicitly flagged as PHANTOM (COVID paper, not cochlear LNP). ✅
  • notes/Prime Editing for STRC.md: Gao 2020 flagged as PHANTOM with correction text. ✅ (h07 scope — not our domain, confirming status)
  • No script or note was found with Gao 2020 PMID 32493791 cited as a real cochlear source.
  • _archive/STRC Hypothesis Ranking Log Archive 2026-Q1.md: historical mentions only, not cited as authoritative.

The h06 and h07 copy-paste phantom is documented, corrected, and not reinstated in any active file.

Summary table

ItemParameterStatusUpgrade
1HILL_K = 200 (K_M)NOT MEASURED — scenario-scan only⚠ stays ⚠, WARNING added to scripts
2HILL_N = 2 (cooperativity)NOT MEASURED — scenario-scan only⚠ stays ⚠, WARNING added to scripts
3MAX_BOOST = 3.0NOT MEASURED — scenario-scan only⚠ stays ⚠, WARNING added to scripts
4STRC_HL_D = 14 dAnchored OOM (Zhang 2012 PMC3267870; Hartig 2024 PMC11459194) — Hartig author corrected⚠ → ✅
5RBM24_HL_D = 2 dAnchored OOM (Schwanhäusser 2011 BioNumbers 106377)⚠ → ✅
6LNP OHC efficiency ≤ 20%No primary paper. Aspirational. Closest: Gao 2018 PMC5784267 ~0.92% (lipid-RNP)⚠ stays ⚠, documented as aspirational
7Gao 2020 PHANTOMConfirmed removed from all scripts✅ clean

h04 subsection

h04 inherits all h06 Hill constants and half-life parameters via shared JSON (rbm24_mrna_dose_response_results.json) and the mrna_lnp_pkpd_integration*.py scripts. The resolution above applies equally to h04. Additionally:

  • strc_mrna_strategy_b_pkpd.py uses LNP_OHC_TARGETED = 0.20 aspirationally (consistent with the tropism scan). No phantom citations found in this script.
  • h04 LNP tropism gap (Item 6) documented above applies.
  • No h04-unique phantom citations found in full vault search.

Ranking delta

#6 mRNA Therapy: B held. No axis change from this audit. Items 4 and 5 are upgrades (⚠→✅) but do not change mechanism/delivery/Misha-fit scores — these were already captured as OOM estimates. Items 1-3 (Hill circular fits) and Item 6 (LNP OHC aspirational) remain open as model limitations. The architectural ceiling finding from STRC h06 Hill Sensitivity Sweep 2026-04-23 is independent of these parameter gaps — the ceiling is math, not Hill-fit dependent. B tier and Mech 2 / Deliv 2 / Misha-fit 2 hold.

#4 mRNA-LNP: B held. No axis change. Strategy B’s hill-free linear dose-response means Item 1-3 gaps don’t affect Strategy B directly; Item 6 (LNP OHC efficiency) is the binding constraint for Strategy B and remains correctly flagged as aspirational.

Connections

Graph View

Backlinks