STRC Research

STRC STRCP1 GTEx Expression Check

7 min read

STRC STRCP1 expression in GTEx — co-expressed at 4.5:1 with STRC, not silent

Question

STRC PE Phase4 STRCP1 Paralog Off-Target killed every PE3b pegRNA on STRCP1 paralog hits. STRC ASO Phase2 STRCP1 Paralog Cross-Hybridization killed every Phase-1 splice-switch ASO on the same paralog. Both atomic notes flagged a shared escape path: if STRCP1 is transcriptionally silent in OHCs, the off-target has no RNA substrate and the paralog risk drops. This proof asks: is that escape path biologically plausible?

GTEx is the cheapest first read. It covers 54 human tissues but no cochlea / inner ear, so we use highest-STRC tissues (cerebellum, testis, brain regions, skin) as proxies. Logic: if STRCP1 is co-expressed with STRC in tissues where STRC is abundant, then the OHC (the highest-STRC tissue in the body) almost certainly co-expresses STRCP1 too.

Method

ohc_strcp1_expression_check.py — single-shot GTEx v8 API query for both genes:

GeneGENCODE v26 IDGTEx tissues
STRCENSG00000242866.954
STRCP1ENSG00000166763.754

Median TPM per tissue pulled via gtexportal.org/api/v2/expression/medianGeneExpression?gencodeId=...&datasetId=gtex_v8. Verdict gate computed as the median STRC:STRCP1 ratio in the top 5 STRC-expressing tissues:

Median ratioVerdictImplication
no STRCP1 detected (<0.05 TPM)STRCP1_SILENT_PROBABLEEscape path PLAUSIBLE for both PE+ASO
≥10:1STRCP1_LOW_FAVORABLEASO loses ~10% dose, tolerable
3:1–10:1STRCP1_CO_EXPRESSED_PARTIALASO loses 25-50% dose, significant
<3:1STRCP1_CO_EXPRESSED_KILLASO loses >50% dose, kills splice ROI

Result — STRCP1 transcribed in all 54 GTEx tissues, ratio 4.5:1 in STRC-expressing tissues

Verdict: STRCP1_CO_EXPRESSED_PARTIAL. Median STRC:STRCP1 ratio across top-5 STRC-expressing tissues = 4.5:1.

Top 10 by STRC TPM (proxy for OHC where direct data is missing):

TissueSTRC TPMSTRCP1 TPMSTRC:STRCP1 ratio
Brain Cerebellum15.172.565.93
Brain Cerebellar Hemisphere12.342.494.96
Testis4.941.114.46
Brain Cortex1.210.383.17
Pituitary1.030.432.38
Brain Frontal Cortex BA90.850.382.23
Skin Sun Exposed0.580.521.11
Brain Nucleus Accumbens0.570.272.11
Skin Not Sun Exposed0.560.451.26
Prostate0.430.361.21

Distribution:

  • STRC > 0.1 TPM: 40 / 54 tissues
  • STRCP1 > 0.1 TPM: 37 / 54 tissues
  • Both > 0.1 TPM: ~36 / 54 tissues (essentially co-expressed wherever either is detected)
  • Tissues where STRCP1 is undetectable (TPM = 0): 0 / 54

The pattern is unambiguous: STRCP1 is not silent. It transcribes wherever STRC does. The Ensembl biotype transcribed_unprocessed_pseudogene and the NCBI note (“a previously suppressed RefSeq was removed because it is now thought that this locus is transcribed”) corroborate independently.

Why this matters more than typical paralog-expression questions

For a typical processed pseudogene, “is it transcribed?” is binary — most processed pseudogenes are not transcribed at all and the off-target paralog risk is moot. STRCP1 is an unprocessed pseudogene retaining the parent’s intron-exon architecture (30 exons vs STRC’s 29; the same splice sites that killed every ASO design). Unprocessed pseudogenes far more often get co-transcribed by the same regulatory elements as the parent gene.

GTEx confirms this empirically. In every STRC-expressing tissue, STRCP1 is co-expressed at 1:1 to 6:1 STRC:STRCP1 ratio. This is not “rare leaky transcription” — STRCP1 reaches 2.6 TPM in cerebellum, comparable to many protein-coding genes there.

The OHC is the body’s highest-expression-per-cell-type tissue for STRC. By extrapolation:

  • If OHC shows the same pattern as cerebellum (the highest-STRC GTEx tissue), STRCP1 expression in OHC is ~2-3 TPM, roughly 16% of the STRC level.
  • If OHC has STRC-supremacy similar to testis/brain (mean ratio 4.5:1), then ~18% of total STRC + STRCP1 mRNA pool is STRCP1.
  • An RNase-H1 gapmer ASO with ≤2 mm match to STRCP1 will cleave 50-80% of STRCP1 transcripts as a function of single dose — dose loss to the paralog is therefore in the 10-20% range, not the >50% range.

This is not the worst-case scenario. The earlier worry that “every ASO molecule binds STRCP1 first” is overstated by the data. But it is also not free — the splice modulation hypothesis was already rated B because of weak inner-ear delivery; losing another ~15% to paralog cleavage is a real (if non-fatal) ROI hit.

For PE the calculus is different. Even rare STRCP1 editing events create permanent off-target genotypes. Unlike ASO (which cleaves transient mRNA and is dose-dependent), PE permanently writes a sequence change. So the GTEx-based dilution argument doesn’t apply to PE — every STRCP1-expressing OHC is at risk for an off-target edit, with unknown phenotypic consequence.

Implications for escape paths

The five escape paths from STRC ASO Phase2 STRCP1 Paralog Cross-Hybridization:

  1. ✗ STRCP1 silent in OHC — falsified by GTEx. STRCP1 transcribed in every STRC-expressing tissue at non-trivial levels. Cannot rely on this.
  2. ✓ Morpholino chemistry (no RNase-H cleavage) — more attractive now. Steric block on STRCP1 splicing has no protein-level consequence (STRCP1 makes no functional protein anyway), and morpholino doesn’t deplete drug at the target. The 4.5:1 GTEx ratio means morpholino dose is barely affected by STRCP1 binding.
  3. △ 3’UTR exonic targets with STRC vs STRCP1 divergence — pseudogene 3’UTRs often diverge faster, but this changes mechanism from splice modulation to steady-state knockdown. Different drug.
  4. △ Whole-genome BLAST + high-Tm discrimination — possible, requires re-design at higher Tm windows (24-26 nt).
  5. ✓ Allele-specific gapmer at c.4976 variant directly — also more attractive now. The c.4976A>C variant is on STRC mutant only (STRCP1 has the WT base). 1-bp ΔTm of ~1.5-2.5 °C is borderline for RNase-H discrimination, but a carefully designed 18-nt gapmer with the variant at the central RNase-H cleavage site could achieve 5-10× preference for the mutant allele. This is the most rigorous chemistry-aware path forward.

For PE Phase 3.5 STRCP1-aware redesign: still required. The GTEx data does NOT relax the need to design pegRNAs that discriminate STRC vs STRCP1 in the seed region.

Limitations

  • No direct OHC data. GTEx covers 54 tissues but not cochlea / inner ear. The proxy assumption (“OHC behaves like cerebellum / testis for STRCP1 ratio”) is conservative but unverified. Direct human OHC RNA-seq would resolve. Closest published: Hoa et al. 2023 PNAS mouse cochlea atlas (PMC10293812) — uses mouse, and their paper does not mention Strcp1 (mouse may lack the ortholog or it was filtered as a pseudogene from standard scRNA-seq pipelines).
  • GTEx v8 = bulk RNA-seq. Median TPM at the tissue level mixes many cell types; per-cell ratio in OHC could differ substantially. Some tissues (cerebellum) have many neuronal subtypes; STRCP1 might be enriched in specific cells while STRC is in others.
  • One snapshot, no developmental data. STRCP1 may have age-dependent expression; if it is high in embryonic OHC and silent in postnatal OHC (or vice versa), the gate decision depends on therapeutic window timing.
  • Allelic balance not assayed. STRCP1 has known alleles with varying integration into the STRC genomic region; some haplotypes may have higher / lower STRCP1 expression. Misha-specific genotype not analyzed here.
  • TPM ≠ functional substrate availability. mRNA stability, accessibility, and ribosome occupancy may differ between STRC and STRCP1; functional substrate for ASO-mediated cleavage tracks bound mRNA, not total TPM.

Files

  • Driver: ~/STRC/models/ohc_strcp1_expression_check.py (stdlib + urllib only, no compute)
  • Output JSON: ~/STRC/models/ohc_strcp1_expression_check.json
  • GTEx API: gtexportal.org/api/v2/expression/medianGeneExpression?...&datasetId=gtex_v8

Ranking delta

  • STRC ASO Exon Skipping: no tier change (stays C). Mechanism axis unchanged. Evidence depth +1 (escape path 1 falsified, escape paths 2 and 5 promoted to “more attractive”). Next step refined from “verify STRCP1 OHC expression first” → “design morpholino splice-switch + allele-specific gapmer at c.4976 in parallel as Phase 3” (escape paths 2 and 5 from the Phase 2 paper). The OHC RNA-seq direct verification is desirable for IND but no longer the gating step — even if STRCP1 is 50% silent in OHC, the GTEx baseline shows we cannot count on full silence.
  • Prime Editing for STRC: no tier change (stays B). Mechanism still 4/5. Evidence depth +1. The STRCP1 paralog off-target stays a per-cell risk regardless of dose dilution — escape path “STRCP1 silent in OHC” was never directly applicable to PE (PE writes permanent edits, not dose-dependent transient cleavage). Next step unchanged: Phase 3.5 STRCP1-discriminating pegRNA redesign is required.
  • All other S/A/B/C tier hypotheses: no change.

The shared “OHC STRCP1 expression check” gate from the prior two notes is now closed (with negative result). Both PE and ASO must proceed via design-side escape paths, not via biology-permission. The cross-cutting next step from the inventory (“ohc_strcp1_expression_check.py”) is now complete — no follow-up RNA-seq scrape needed unless a wet-lab RNA-FISH experiment is later proposed.

Connections

Graph View

Backlinks