STRC h05 Ca Oscillation — Parameter Provenance Audit 2026-04-23
Post-audit fix of ca_oscillation_ac1_creb_pivot.py and ca_oscillation_rbm24_ode.py. Phase 2 / Phase 3 / maternal-only scripts scanned — no additional phantom cites found there. Companion: calcium-oscillation. Triggered by STRC Cross-Hypothesis Parameter Audit 2026-04-23 flag “3 phantoms + 2 value-citation mismatches”.
Phantom citations — REMOVED and flagged
ca_oscillation_ac1_creb_pivot.py
1. “Wu et al. 2011” (AC1 kinetics) — REMOVED. No such paper exists. The AC1 K_Ca = 150 nM and Hill n=2 values are replaced with references to:
- Willoughby & Cooper 2007 Physiol Rev (PMID 17615394): Ca EC50 range 100-500 nM, Hill 1.5-2.5.
- Masada et al. 2012 Biochemistry (PMID 22971080): mechanistic basis for Ca/CaM activation.
The 150 nM EC50 value is within the published range; Hill n=2 is within 1.5-2.5. But the AC1 Vmax = 2000 nM/s (120 µM/min) is ~30-60× higher than the published ~2-5 µM/min in membrane prep — flagged as cell-volume-normalisation choice, not a measurement.
2. “Sharma 2018” (STRC mRNA splicing t½ = 2 h) — REMOVED. No such paper exists. STRC_MRNA_DECAY_S = 1e-4 retained as a model parameter flagged ⚠ UNSOURCED. Note: ca_oscillation_rbm24_ode.py uses k_mRNA_deg = 0.0004 (t½ ≈ 30 min) — 4× faster than pivot script. Cross-script inconsistency noted in both scripts.
3. “Visel 2006” — removed as unchecked companion cite to Wu 2011.
4. “Cha et al. 2010” (CRE promoter Hill n=1, K_half 0.2) — REMOVED. PMID never verified. K_CREB_CRE_HALF = 0.2, N_CRE = 1.0, K_TXN_MAX_FOLD = 6.0 retained as ⚠ UNSOURCED model parameters.
5. “Houslay 2010” (PDE4 Vmax, Km) — WEAKENED to “Houslay lab reviews (no single PMID confirmed)“. Values kept in range; flagged ⚠ NEEDS SPECIFIC PMID.
ca_oscillation_rbm24_ode.py
6. “Krey 2015” (target_protein = 15,000 molecules/OHC) — REMOVED. Krey et al. 2015 Scientific Data (Wilmarth et al. stereocilia proteome, PMID 25977827) does NOT quantify STRC copy number per OHC. No paper anywhere quantifies STRC per OHC (audit 2026-04-23). Retained as ⚠ UNSOURCED calibration set-point with explicit comment.
7. “Chao et al. 2010” (CaMKII) — CORRECTED to Chao 2011 Cell (PMID 21458670). Year was likely typo. Value order-of-magnitude plausible but not an exact measurement — flagged.
Value-citation mismatches — reconciled and flagged
(A) PKA K_cAMP = 300 nM vs “Zaccolo 2007 = 100 nM” — reconciled. The 300 nM value is correct for in-cell PKA activation (Surdo et al. 2017, PMID 29074866, in-cell FRET). In-vitro PKA K_cAMP is ~100 nM (Zaccolo & Pozzan 2002); in-cell allosterics raise it to 300 nM. Docstring now cites Surdo 2017.
(B) CREB-P dephosphorylation k = 0.005/s vs Gonzalez & Montminy 1989 implied k = 0.0012-0.0023/s — model runs 2-4× faster than literature. This speed-up makes pCREB more transient and affects the 1.82× silence→100 dB fold-change. No primary source for the faster rate in hair cells. Flagged in docstring + comment as ⚠ model choice.
(C) STRC mRNA t½: 2 h (pivot) vs 30 min (rbm24) — 4× cross-script disagreement. Both unsourced; no primary STRC mRNA stability data exist. Both scripts now carry the cross-reference warning.
(D) STRC protein t½: 38 h (pivot) vs ~30 days (rbm24) — 20× cross-script disagreement. Both unsourced. rbm24’s 30-day figure anchored to unrealistic “Krey 2015 = 15,000 per OHC” target. Resolution requires primary STRC degradation data (not in literature).
Post-fix script outputs
ca_oscillation_ac1_creb_pivot.py
Runs clean. Output verdict unchanged: “HYPOTHESIS SUPPORTED — AC1-CREB pathway gives monotonic Ca→STRC up-regulation (1.82× silence→100dB). CREB-P fold 47.7×. Saturates above ~45 dB.”
Caveat now explicit in script: the 1.82× fold-change depends on the 2-4× faster dephosphorylation rate than literature. A literature-consistent dephos rate would produce a larger fold-change (stronger argument for the hypothesis), but the absolute number is anchored in unverified parameters.
ca_oscillation_rbm24_ode.py
Runs clean. Maternal-allele rescue gate numbers unchanged, but now carry:
- target_protein = 15,000 UNSOURCED flag
- k_prot_deg t½ ~30 days ⚠ fitted flag with cross-script warning
Phase 3 bifurcation — still robust
The topological result (limit cycle exists for the 7-variable AC1-PKA-CREB ODE system at certain parameter combinations) does not depend on exact parameter values, only on their qualitative structure (AC1 Hill > 1, PDE4 saturating, CREB positive feedback via STRC → bundle stiffness → Ca). Phase 3 topological verdict stays: limit cycle exists, requires CREB-P feedback.
Ranking delta
| Axis | Pre-audit | Post-audit | Reason |
|---|---|---|---|
| Mech | 3 | 3 | Topological result unchanged; quantitative claims now flagged but the mechanism sketch is intact |
| Deliv | 4 | 4 | Acoustic therapy delivery path unchanged (doesn’t depend on phantom parameters) |
| Misha-fit | 2 | 2 | Unchanged — Misha post-onset, deafened; Ca oscillation hypothesis doesn’t rescue structural HTC loss |
New tier: A (unchanged). Audit outcome: quantitative claims (fold-changes, absolute protein counts) need “UNSOURCED” asterisks; topological / qualitative claims defensible. Not a tier demotion because the hypothesis’s strategic weakness is Misha-fit (2), not the computational fidelity.
Next steps before promotion to S: measure STRC mRNA / protein t½ in an OHC line (wet); measure AC1 expression in OHCs (snRNAseq or IHC); compute-side, no new work unlocks S without wet data.
Connections
[part-of]STRC Hypothesis Ranking[part-of]STRC Cross-Hypothesis Parameter Audit 2026-04-23[see-also]STRC Calcium Oscillation Acoustic Therapy[see-also]calcium-oscillation[about]Misha