STRC E1659A — Conservation Analysis and VUS Reclassification
The Variant
- Gene: STRC (stereocilin), UniProt Q7RTU9
- Variant: c.4976A>C, p.(Glu1659Ala) — E1659A
- Chromosome: 15q15.3, GRCh38 position chr15:43600551
- WES report: Hong Kong Children’s Hospital, Lab No 23C7500174 (Dec 2022)
- Was: VUS (Variant of Uncertain Significance)
- Was reclassified to LP (2026-03-16), but corrected to VUS-high (2026-04-01) after PP1 misapplication identified by Jeffrey Holt
Conservation Analysis (New Evidence, 2026-03-15)
E1659 is 100% conserved across 9 mammalian species. Egor extracted aligned STRC sequences from UniProt and aligned them manually with Claude.
| Species | Position 1659 | Surrounding motif |
|---|---|---|
| Human | Glu (E) | PEIFTEIGTIAAG |
| Mouse | Glu (E) | PEIFTEIGTIAAG |
| Rat | Glu (E) | PEIFTEIGTIAAG |
| Cow | Glu (E) | PEIFTEIGTIAAG |
| Monkey | Glu (E) | PEIFTEIGTIAAG |
| Pig | Glu (E) | PEIFTEIGTIAAG |
| Dog | Glu (E) | PEIFTEIGTIAAG |
| Bat | Glu (E) | PEIFTEIGTIAAG |
| Bear | Glu (E) | PEIFTEIGTIAAG |
Identical motif across species: PEIFTEIGTIAAG (the surrounding 13-aa motif is also conserved)
~80 million years of evolutionary conservation. This is strong evidence of functional importance: if any change were tolerated, evolution would have explored it.
Conservation Footnote
The reason standard bioinformatics tools (SIFT, PolyPhen-2, CADD) often fail on STRC variants: the STRCP1 pseudogene on chr15q15.3 has ~99.6% nucleotide identity to STRC. Alignment tools confuse STRC and STRCP1 sequences, artificially inflating the apparent “variation” at STRC positions. AlphaMissense works because it uses protein structure, not DNA sequence alignment. See STRC Pseudogene Problem for full explanation.
ACMG/AMP Classification Framework (2015)
Criteria Applied
| Code | Evidence | Level | Justification |
|---|---|---|---|
| PM2_Supporting | Absent from gnomAD | Supporting | 0 alleles in 251,000+ gnomAD controls |
| PM3 | In trans with pathogenic | Moderate | Found opposite confirmed pathogenic 98kb deletion (MLPA, allele 1) |
| PP1_Supporting | Evolutionary conservation | Supporting | 100% conserved E1659 in 9 mammals, ~80M years |
| PP3_Moderate | Computational evidence | Moderate | AlphaMissense 0.9016 + REVEL 0.65, concordant pathogenic predictions |
Combined: 2 Moderate + 2 Supporting = Likely Pathogenic (per Richards et al. 2015 ACMG guidelines and Pejaver et al. 2022 computational evidence calibration)
Previous Classification (Before 2026-03-16)
Had only: PM2_Supporting (absent from gnomAD) + PM3_Moderate (in trans) = 1 Moderate + 1 Supporting = VUS
New Evidence Added
- PP1_Supporting: Conservation across 9 species (Egor’s analysis, cross-referenced with UniProt alignment data)
- PP3_Moderate upgrade: AlphaMissense 0.9016 qualifies for Moderate level per Pejaver 2022 (threshold for Moderate: 0.840)
AlphaMissense Saturation Analysis
Source: alphamissense_heatmap_uniref90_2022.csv (Google DeepMind, all human proteins)
- AlphaMissense score for E1659A: 0.9016 (scale: 0 = benign, 1 = pathogenic)
- Classification: Likely Pathogenic (>0.564 = pathogenic threshold; >0.840 = Moderate per Pejaver 2022)
- Mean AlphaMissense score for all STRC positions: 0.714
- Fraction of STRC positions predicted LP: 68.3%
- Position 1659 is in the top quartile of pathogenicity scores for STRC
What AlphaMissense Tested (Saturation Mutagenesis)
AlphaMissense evaluated every possible amino acid change at every position. For position 1659:
- E1659A: 0.9016 (Likely Pathogenic)
- E1659K: 0.92 (any positive charge introduction = even worse)
- E1659Q: 0.81 (conservative glutamine substitution = moderate)
- E1659D: 0.23 (benign — Asp is also negatively charged, partially functional)
The fact that E→D is benign while E→A is pathogenic confirms: it is the charge at position 1659 that matters, not the size. Glutamate provides a -1 charge at pH 7.4. Alanine provides 0. This is the basis of the electrostatic analysis.
Step 6 of Reclassification Process
Full methodology is documented in STRC Research Methodology. Step 6 specifically involved:
- Downloading STRC protein sequences for 9 species from UniProt
- Running multiple sequence alignment with ClustalOmega (via Claude/API)
- Identifying the PEIFTEIGTIAAG motif as 100% conserved
- Mapping to ACMG PP1_Supporting criterion
- Checking REVEL score (0.65, obtained from ClinVar/dbNSFP)
- Verifying AlphaMissense score against Pejaver 2022 thresholds
- Recalculating ACMG total: VUS → Likely Pathogenic
CRITICAL CORRECTION (2026-04-01) — PP1 Misapplication
Per Jeffrey Holt and his geneticist colleague (March 31 reply):
PP1 (co-segregation) was incorrectly applied in the original reclassification. PP1 requires co-segregation evidence from affected family members with hearing loss. No affected relatives with hearing loss exist in Egor’s family, so PP1 CANNOT be applied. Evolutionary conservation is already captured inside PP3 (in silico evidence). Using conservation as PP1 = double-counting.
Corrected Classification (without PP1):
| Code | Evidence | Level |
|---|---|---|
| PM2_Supporting | Absent from gnomAD | Supporting |
| PM3 | In trans with pathogenic | Moderate |
| PP3_Moderate | AlphaMissense 0.9016 + REVEL 0.789 | Moderate |
Combined: 2 Moderate + 1 Supporting = VUS (does not reach Likely Pathogenic threshold)
Holt Lab Framework: VUS-high
Jeff’s lab uses a subclassification system. E1659A scores VUS-high, meaning:
- Not enough evidence for LP by strict ACMG rules
- But “for clinical trials, many sites might be willing to allow enrollment with VUS-high, especially for recessive disease with one allele P/LP”
Additional novel evidence (from Jeff):
- E1659A not found in seqr (70,000 cases worldwide) or Franklin (thousands of clinical labs) — truly novel
- This extreme rarity further supports pathogenicity but is already captured in PM2
HKCH Status (2026-04-01):
- Reclassification request submitted (VUS → LP) — HKCH called back: not enough evidence
- Genetics panel meets once a year — next: November 2026
- CAN accept new evidence submissions before November
- Action: need additional evidence lines (functional data, more in silico tools, or family segregation data if available)
What’s Needed to Reach LP:
One more Moderate-level criterion would do it:
- PS3_Supporting (functional evidence) — lab testing of E1659A effect on stereocilin
- PP4 (phenotype specificity) — if SNHL pattern matches STRC-specific audiometric profile
- Better calibrated computational tools reaching Strong level
Clinical Impact of Reclassification
Reclassifying from VUS to Likely Pathogenic:
- Clinical trial eligibility: most gene therapy trials require confirmed pathogenic variants (not VUS)
- France MDPH disability: Likely Pathogenic strengthens the application vs VUS
- HK Children’s Hospital records: formal request sent 2026-03-15 to update Misha’s file
- Reproductive counseling: confirmed biallelic diagnosis clarifies recurrence risk
Reclassification Letters Sent
- HK Children’s Hospital Genetics: Formal reclassification request (VUS → Likely Pathogenic), 2026-03-15
- Documents:
~/Documents/Disability Card/misha-france-mdph/LETTER-Reclassification-Request-STRC-2026-03-15.pdf
- Documents:
Scientific Documentation
- Full analysis PDF:
~/Documents/Disability Card/misha-france-mdph/STRC-E1659A-Full-Analysis-2026-03-15.pdf - Research site: https://strc.egor.lol
Connections
- STRC Hearing Loss — variant classified as VUS-high (corrected from initial LP reclassification)
[supports]STRC Electrostatic Analysis E1659A — why E1659A is pathogenic (charge lost)[supports]STRC Pseudogene Problem — pseudogene explains why standard tools failed[see-also]STRC AlphaFold3 Computational Experiments — Job 3 shows no structural damage[see-also]STRC Research Methodology — how Egor performed the reclassification step-by-step[see-also]Prime Editing for STRC — reclassified variant is the therapeutic target- Misha-Hearing-10-Year-Plan
[see-also]France MDPH Disability Benefits[about]Misha