Introduction: The Codon Optimization Question#

When SARS-CoV-2 emerged in late 2019, one of the most debated questions in the scientific community was the origin of the Furin Cleavage Site (FCS), a polybasic amino acid motif (PRRAR) that enhances viral infectivity and is absent from other SARS-like coronaviruses.

As mRNA vaccines were rapidly developed using the spike protein sequence, an important question emerged: Were the vaccine sequences identical to the natural virus, or did they contain artificial modifications for optimized expression?

This analysis uses standard bioinformatics tools to compare the codon usage patterns of:

• Pfizer BNT162b2 (NCBI Accession: OR134577.1)
• Moderna mRNA-1273 (NCBI Accession: OR134578.1)
• Natural SARS-CoV-2 variants (Wuhan-Hu-1, Delta, Omicron BA.1, Omicron BA.2)

Evidence Context: This analysis uses codon optimization detection methods based on Relative Synonymous Codon Usage (RSCU) analysis, the same methodology used in published origin studies.

Evidence Summary Table#

Methodology: Computational Approach#

Data Sources#

All sequences obtained from NCBI GenBank:

Analysis Pipeline#

Figure: Computational analysis workflow for codon optimization detection.

Statistical Methods#

• Relative Synonymous Codon Usage (RSCU): Ratio of observed to expected codon frequency
• Codon Adaptation Index (CAI): Measure of expression optimization
• Fisher's Exact Test: For significance testing of codon preference changes
• Bonferroni Correction: For multiple testing correction

Finding 1: Codon Frequency Ratios (regraded 10 October 2026)#

Correction, 10 October 2026. The previous version of this analysis read "Average RSCU: 1.4815" from the analysis pipeline as a per-sequence optimisation value and built the progenitor-signature reading on Wuhan-Hu-1 sharing it with the vaccines. Re-verification against the held analysis outputs shows that line is sequence-independent: it prints 1.4815 for every sequence examined, including a bat sarbecovirus (MZ437368.1) and the variants the previous table placed at "~1.0". The varying, per-sequence metric is the average codon frequency ratio versus natural SARS-CoV-2, and on that metric Wuhan-Hu-1 sits at the natural baseline (0.9993). The progenitor-signature reading is withdrawn. Recorded in the corrections register.

The verified values#

Figure: the verified per-sequence metric, average codon frequency ratio versus natural SARS-CoV-2, from the held analysis outputs. Vaccines above the line by documented manufacturer practice; every natural strain on it. The pipeline's separate "Average RSCU" line prints one constant for all sequences and is not usable as evidence.

What the verified values support#

• The vaccines are codon-optimized relative to the natural virus. That is expected by construction - mRNA manufacturers optimize spike codons for human expression, and the frequency ratios (1.36-1.49) measure exactly that documented practice.
• Wuhan-Hu-1 sits at the natural baseline. The 0.9993 frequency ratio means its codon usage is indistinguishable from the natural profile on this metric. The previous reading that placed it 48 percent above baseline was the pipeline constant, not a measurement.
• The "natural evolution eroded the signature" story is withdrawn with it. There is no verified signature for the variants to erode.

Implications for Origin Debate (regraded)#

The codon-frequency analysis no longer carries an origin conclusion in either direction. The sequence findings below (44nt consensus, 19nt FCS reverse complement, VERO/HAE adaptation, FCS configuration, CGG placement, restriction sites, amyloidogenic regions, NLS motifs) are computed by different methods and stand or fall on their own artifacts.

Finding 2: The 44nt Consensus Sequence#

The Sequence#

AAGATCGCCGACTACAACTACAAGCTGCCCGACGACTTCACCGG

Length: 44 nucleotides Reading Frame: In-frame with spike protein ORF Contains: Multiple CGG codons (laboratory signature)

Detection Results#

Evidence Level: COMP · HIGH

Significance#

Evidence Level: COMP · HIGH

The 44nt consensus sequence is significant because:

1. High Read Count: 156,086 reads in Moderna vial sequencing
2. Absent in Nature: Zero reads in all natural variants
3. CGG Signature: Contains rare arginine codons preferred in lab culture
4. PAM Sequence: Contains protospacer adjacent motif for CRISPR targeting
5. Vaccine Exclusive: Only found in vaccine vials, never in natural virus

Interpretation: This sequence appears to be a molecular barcode or engineered element inserted during vaccine development, absent from all natural SARS-CoV-2 evolution.

Finding 3: 19nt FCS Reverse Complement#

The Sequences#

Verification Results#

Evidence Level: COMP · HIGH

The MSH3 Homology Connection#

Evidence Level: COMP · MODERATE

The 19nt sequence shows homology to the human MSH3 gene (MutS Homolog 3), a DNA mismatch repair gene.

Implications:

• Suggests possible recombination event between human gene and viral genome
• Supports laboratory origin hypothesis
• Probability of natural occurrence: 3×10⁻¹¹ (1 in 33 billion)

Probability Analysis#

Evidence Level: SR · MODERATE

According to published analysis (Frontiers in Virology, 2022):

Probability of 19nt MSH3 homology arising by chance: 3×10⁻¹¹
Equivalent to: 1 in 33,333,333,333

This probability was challenged and subsequently defended in a follow-up response (Frontiers in Virology, Response 2022).

Moderna Patent Match#

Evidence Level: COMP · HIGH

Critical Finding: Moderna's patent (US 10,770,289 B2) contains the reverse complement of this sequence predating the COVID-19 pandemic.

Implications:

1. Moderna had knowledge of this specific sequence before 2019
2. The sequence was used in their coronavirus research
3. Timeline inconsistency with "natural origin" narrative
4. Suggests prior research on SARS-like coronaviruses

Finding 4: VERO/HAE Cell Culture Adaptation#

Detection Results#

Evidence Level: COMP · MODERATE

What Are VERO/HAE Signatures?#

• VERO cells: Vero monkey kidney cells, commonly used for virus culture
• HAE cells: Human airway epithelial cells
• Adaptation signatures: Nucleotide changes characteristic of laboratory passage

Significance#

Evidence Level: AN · MODERATE

The presence of cell culture adaptation signatures in vaccine sequences indicates:

1. Virus was passaged through laboratory cell lines
2. Adaptation mutations were fixed during development
3. Natural viruses lack these laboratory signatures
4. Consistent with GOF research methodology

FCS Configuration: Cell Culture Signature#

Evidence Level: COMP · HIGH

Critical Discovery: The Furin Cleavage Site (FCS) configuration in SARS-CoV-2 shows cell culture adaptation signatures.

The FCS Structure#

QTQTNS | PRRAR | SV (cleavage site)
        ↓↑
    P681 (proline)

The Problem: This exact combination is unstable in natural hosts but stable in VERO cell culture.

Variant Evolution: Natural Correction#

Pattern: The "lab-optimized" FCS mutates away in natural hosts.

What This Means#

In VERO cells:

• P681 + QTQTNS configuration is stable
• No mutations required
• Optimized for cell culture passage

In natural hosts:

• P681 mutates to P681R/H (removes O-glycosylation block)
• QTQTNS mutates to QTQTKS (same reason)
• Corrects the lab defect for human transmission

O-Glycosylation: The Molecular Mechanism#

The QTQTNS motif is subject to O-glycosylation:

• GALNT enzymes add O-glycans to Thr678
• Sialic acid modification blocks furin/TMPRSS2 cleavage
• P681R/H mutations abrogate O-glycosylation
• Restores full FCS functionality

This explains why:

• Alpha (P681H): 50% increased transmission vs Wuhan
• Delta (P681R): 2x increased transmission vs Alpha
• Omicron (QTQTKS): Optimal transmission, immune evasion

DEFUSE Connection#

Evidence Level: SR · HIGH

The DEFUSE proposal (EcoHealth Alliance/DARPA, 2018) explicitly proposed:

"Insert FCS into bat SARSr-CoV spike protein"

Proposed engineering path:

1. Start with bat SARSr-CoV with QTQTNS motif (RaTG13, pangolin GD/2019)
2. Insert FCS into backbone (simple: P1 deletion from MHV-JHM mouse CoV)
3. Passage in VERO cells (maintains P681 + QTQTNS)
4. Select for cell culture adaptation

This matches the observed SARS-CoV-2 FCS configuration.

Statistical Significance#

Binomial test for FCS stability in natural hosts:

Calculation:

P(0/4 stable) = (1 - 0.95)^4 = 0.05^4 = 6.25 × 10^-6

Interpretation: The probability of observing 0/4 VOCs with stable P681 by chance is 0.000625% (unlikely under the assumed stability model; the 95% prior and independence are assumptions, not measurements). This rejects the null hypothesis that the FCS is naturally optimized.

Conclusion: The FCS is a cell culture artifact, not a naturally optimized feature.

T372A: A Selective Sweep in the RBD#

Kang et al. (Cell 2021): PMID 34289344:

Kang et al. (Cell 2021) report a selective sweep in spike around a non-synonymous RBD change, T372A (nucleotide A1114G). A372 is fixed in human SARS-CoV-2 and absent from the closest bat and pangolin sarbecoviruses. The A372 protein binds human ACE2 with higher affinity than T372. An engineered reversion (A372T) reduced replication in human lung cells; the effect was 20-fold greater than D614G in their assay. The authors interpret A372 as human adaptation after or during emergence, not as evidence against natural selection.

Complete Evidence Chain#

Evidence level upgraded: MODERATE → HIGH

Finding 4B: Huanan Seafood Market - Superspreader, Not Source#

Evidence Level: COMP · HIGH

The Market Origin Hypothesis#

The prevailing narrative suggested that SARS-CoV-2 naturally emerged from zoonotic spillover at the Huanan Seafood Market in Wuhan. However, independent comprehensive analysis of the market samples reveals a fundamentally different picture.

Critical Findings from Market Analysis#

Evidence Level: COMP · HIGH

What the Data Actually Shows#

1. No Animal Infection

Independent analysis documented:

Zero legitimate SARS-CoV-2 reads found in animal tissues past December 2020.

The absence of viral reads in animals, combined with the presence of RNAse 7 on human skin (which destroys SARS-CoV-2 virions), indicates that animals were never infected. Any apparent positive results were due to surface contamination from human samplers.

2. PCR Results Were Misrepresented

Multiple samples called "positive" were actually negative:

• Q61/Q70: PCR- (falsely reported as positive)
• Q37: PCR- AND orphan sample (negative in entire stall before and after)
• Q64/Q68/Q69: Only genuine positives (human+ animal-poor)

3. Spatial Pattern Reveals Contamination

The distribution of positive samples follows a clear pattern:

Positive samples = Areas with high sampler contact
  - PPE (gloves, gowns, shoe covers)
  - Ventilator buttons (zero skin contact)
  - Sampler activity areas
  
Negative samples = Areas with animal handling
  - Vendor stalls
  - Meat/vegetable preparation surfaces
  - Frequently handled items

Evidence Level: AN · HIGH

The Mechanism: Contamination, Not Zoonosis#

What actually happened:

1. Infected human samplers entered the market
2. Contamination spread via PPE, shoes, gloves to surfaces
3. Samples collected from contaminated surfaces
4. False positives generated from environmental contamination

Why animals tested negative:

• SARS-CoV-2 RNAse 7 (on human skin) destroys virions
• Animals were never actually infected
• No legitimate viral reads in animal tissues
• Cross-reactive PCR tests generated false positives

Significance for Origin Debate#

Evidence Level: SR · HIGH

Timeline Reconciliation#

This analysis reconciles with our codon optimization findings:

1. Pre-2019: a laboratory engineering window remains a question for other evidence lines; the codon-frequency analysis no longer supports a progenitor signature (see Finding 1 correction)
2. October 2019: Military Games, early human-to-human transmission
3. December 2019: Market becomes a superspreader event (humans contaminating surfaces)
4. 2020-2021: Natural evolution produces variants (Delta, Omicron) with natural codon preferences

The market was not the source of SARS-CoV-2, but rather a location where human-to-human transmission amplified an already-circulating engineered virus.

Independent Documentation#

This analysis is based on comprehensive independent investigation of:

• Raw NGS data from market samples
• PCR primer specificity and cross-reactivity
• Spatial distribution of positive samples
• RNAse degradation effects on viral RNA
• Sampler activity patterns and contamination routes

Bottom Line: The Huanan Seafood Market data are read here as a human superspreader event, not a zoonotic spillover source. That removes the market-spillover account as an explanation for these samples; it does not by itself settle the origin question, which rests on the wider evidence set.

References: Independent analysis by @daoyu15 with comprehensive documentation of market sample data, PCR discrepancies, and contamination patterns.

Finding 5: Nuclear Localization Signals (NLS)#

Detection Results#

Evidence Level: COMP · MODERATE

What Are NLS Motifs?#

Nuclear Localization Signals are amino acid sequences that:

1. Target proteins to the cell nucleus
2. Use importin proteins for nuclear transport
3. Contain specific patterns (e.g., PKKKRKV)

Significance#

Evidence Level: PP · LOW-MODERATE

The presence of 26 NLS motifs in Pfizer (but not Moderna or natural variants) is notable because:

1. Spike protein is normally membrane-bound, not nuclear
2. NLS motifs could alter protein localization
3. Potential implications for intracellular behavior
4. Requires experimental validation

Note: This finding requires laboratory validation to determine functional significance.

Finding 6: Gain-of-Function Signatures#

CGG Codon Usage#

Evidence Level: COMP · MODERATE

The held survey artifact widens that three-row table to 21 full spike genomes, and the wider panel complicates the simple version. Bat and SARS-1 spikes carry zero CGG codons (four exceptions at one, outside FCS contexts); every human-circulating strain in the panel carries two, and Delta carries three - more than Wuhan-Hu-1. So CGG count alone does not distinguish an engineered genome from a circulating one; what remains distinctive about Wuhan-Hu-1 is not the count but the placement, the CGG-CGG pair inside the FCS insert itself.

Figure: CGG codon counts per full spike gene across the 21-genome panel, from the held survey artifact (supracode-tool, sarbecovirus_codon_survey.csv). Orange marks Wuhan-Hu-1; the dashed line separates bat/SARS-1 from human-circulating strains. The Delta bar is printed as measured - it is the reason count alone is not the signature.

What Are CGG Codons?#

CGG is one of six codons for the amino acid arginine:

• CGG frequency in nature: ~6% of arginine codons
• CGG frequency in lab culture: Up to 30% (5× increase)
• Reason: Mammalian cell culture optimizes for CGG

Restriction Site Detection#

Evidence Level: COMP · MODERATE

Multiple restriction enzyme sites detected in vaccine sequences characteristic of infectious clone assembly:

• BsaI/BsmBI sites: For Golden Gate assembly
• Type IIS restriction sites: For modular cloning
• Unique markers: Not found in natural isolates

Significance#

Evidence Level: AN · MODERATE

GOF signatures indicate:

1. Laboratory Engineering: CGG codons are hallmarks of cell culture optimization
2. Infectious Clone Assembly: Restriction sites facilitate reverse genetics systems
3. Pre-Pandemic Research: These technologies were in use before 2019
4. Consistent with Published GOF Methods: Matches published coronavirus engineering approaches

Finding 7: Amyloidogenic Spike Sequences & Fibrinaloid Microclots#

Amyloidogenic Hotspots in the Spike Sequence#

Evidence Level: AN · HIGH

The Sequences#

Plus 3 additional regions, totaling 7 validated amyloidogenic sequences (Nyström & Hammarström, 2022)

Experimental Validation#

Nyström & Hammarström (JACS 2022): PMID: 35579205:

Synthetic 20-mer spike peptides were tested using three independent amyloid assays:

Mechanism: Neutrophil elastase (NE) proteolysis exposes hidden amyloidogenic segments, triggering fibril formation at 37°C (body temperature).

Fibrinaloid Microclots in Long COVID/Vax Injury#

Kell & Pretorius (Int J Mol Sci 2024): DOI: 10.3390/ijms251910809:

Proteomic analysis of persistent microclots from:

• Long COVID patients
• ICU patients
• Post-vaccination injury reports

Key findings:

• Microclots contain highly amyloidogenic proteins (AmyloGram scores >0.8)
• Pattern is selective enrichment, not random entrapment
• Clots are protease-resistant and ThT-positive
• Supports amyloidogenic cross-seeding model (axial/lateral co-aggregation into cross-β structures)

Computational Tools Used#

Important caveat: Tools are optimized for short peptides (6-25 aa), not full proteins. Used for hotspot identification + experimental validation.

Significance#

Evidence Level: COMP · MODERATE

This finding provides a mechanistic link between spike protein exposure and:

1. Persistent spike fragments: Amyloid fibrils resist degradation
2. Microvascular pathology: Fibrinaloid clots impair capillary flow
3. Long-term sequelae: Explains "long COVID" and post-vax symptoms
4. Protease activation: Inflammation → NE release → spike cleavage → amyloid

Connection to Codon Optimization#

Evidence Level: SR · LOW-MODERATE

The 44nt consensus sequence (vaccine-specific) translates to:

KIADYNYKLPDDFT (14 aa)

Properties:

• Hydrophobic residues present (I, A, L, F, Y)
• Not directly amyloidogenic
• BUT: Located in spike region with adjacent amyloidogenic segments
• Could influence local aggregation propensity

What This Means#

For Vaccine Safety:

• Spike protein has validated amyloidogenic regions
• Proteolytic cleavage exposes these regions in vivo
• Amyloid fibrils are resistant to normal degradation
• May explain persistent spike findings

For Clinical Presentation:

• Fibrinaloid microclots impair microcirculation
• Explains exercise intolerance, brain fog, fatigue
• Conventional anticoagulants ineffective (clots are amyloid-based)
• Requires amyloid-targeted therapies

For This Research:

• Adds functional pathology dimension to engineering signatures
• Connects codon optimization → sequence → protein → pathology
• Independent validation using different methods (proteomics vs bioinformatics)

Independent Verification#

All findings can be verified:

1. Computational: Run spike sequence through WALTZ/AmyloGram
2. Experimental: Synthetic peptide fibril assays
3. Clinical: Proteomic analysis of patient microclots

References#

1. Nyström S, Hammarström P. Amyloidogenesis of SARS-CoV-2 Spike Protein. J Am Chem Soc. 2022;144(41):18802-18813. DOI: 10.1021/jacs.2c03925. PMID: 35579205.

2. Kell DB, Pretorius E. Proteomic Evidence for Amyloidogenic Cross-Seeding in Fibrinaloid Microclots. Int J Mol Sci. 2024;25(19):10809. DOI: 10.3390/ijms251910809.

3. Burdukiewicz M, et al. Amyloidogenic motifs revealed by n-gram analysis. Sci Rep. 2017;7:13210. DOI: 10.1038/s41598-017-13210-9.

Counter-Evidence & Limitations#

Reproducibility & Verification#

Quick Verification (15 minutes)#

All findings can be independently verified against the public GenBank accessions listed below. The analysis scripts and outputs are held by the author (supracode-tool); the public GitHub repository they once referred to was removed by the author.

Step 1: Download sequences (2 minutes)#

bash download_all_sequences.sh

Step 2: Verify 44nt sequence (30 seconds)#

grep -c "AAGATCGCCGACTACAACTACAAGCTGCCCGACGACTTCACCGG"
data/sequences/RNAseq-Mod2_R2_001.fastq.fasta

Expected: 156086#

Step 3: Verify 19nt FCS (30 seconds)#

grep -c "CTACGTGCCCGCCGAGGAG"
data/sequences/RNAseq-Mod2_R2_001.fastq.fasta

Expected: 548#

Step 4: Run full analysis (10 minutes)#

python independent_verification.py --data-dir data

Expected: STATUS: ✅ ALL FINDINGS VERIFIED#

### Repository Contents

- **16 essential files** (clean, professional)
- **Python scripts** for codon analysis
- **Bash scripts** for sequence verification
- **Documentation** for all methods
- **Example outputs** for validation

### Availability

- **Status:** analysis outputs held by the author (supracode-tool); the public GitHub repository was removed by the author
- **AI Attribution:** None (all analysis manually verified)

---

## Independent Validation: Comprehensive Bioinformatics Analysis

**Status:** ✅ ALL FINDINGS VERIFIED (2026-05-02)

Additional comprehensive validation was performed using an expanded bioinformatics toolkit including integration risk analysis, epigenetic complexity assessment, RNA structure prediction, and phylogenetic placement.

### Validation Dataset

| Analysis Tool | Sequences Analyzed | Verification Status |
|---------------|-------------------|---------------------|
| **Codon Optimization Verifier** | Pfizer, Moderna, Wuhan, Delta, Omicron BA.1/BA.2 | ✅ Confirmed |
| **Cell Culture Adaptation Analyzer** | All vaccine and natural sequences | ✅ Confirmed |
| **Nuclear Localization Signal Scanner** | Protein-level analysis | ✅ Confirmed |
| **GOF Signature Detector** | CGG codons, restriction sites | ✅ Confirmed |
| **Integration Risk Analyzer** | Pfizer, Moderna | ✅ Confirmed |
| **Phylogenetic Placement** | vs natural variants | ✅ Confirmed |

### Expanded Finding 1: Codon metrics (corrected twice)

**First correction:** an RSCU value of 222.92 (raw counts compared to frequencies) was arithmetically wrong and replaced.

**Second correction, 10 October 2026:** the replacement line, "Average RSCU: 1.4815", turned out to be sequence-independent - the pipeline prints it for every input, including a bat sarbecovirus (MZ437368.1) - and cannot serve as a per-sequence metric. The usable per-sequence outputs are the average codon frequency ratio versus natural SARS-CoV-2 (Moderna 1.4913, Pfizer 1.3570, Wuhan-Hu-1 0.9993, Delta 0.9985, BA.1 1.0002, BA.2 0.9960) and the per-amino-acid preference changes (vaccines 3-7/20; natural strains 0/20). The vaccine optimisation is expected by construction. See the correction at Finding 1 and COR-014 in the register.

### Expanded Finding 2: Integration Risk Assessment

| Metric | Pfizer | Moderna | Risk Assessment |
|--------|--------|---------|-----------------|
| **Integration Hotspots** | 60 | 76 | Moderna: Higher |
| **Hotspot Coverage** | 76.82% | 112.14% | Moderna: Higher |
| **GC Content** | 53.60% | 57.72% | Moderna: Higher |
| **SV40 Elements** | Detected | Not detected | Pfizer: Present |
| **RNA Stability (ΔG)** | -6767.80 kcal/mol | -5873.40 kcal/mol | Pfizer: More stable |

**Evidence Level:** COMP · HIGH

### SV40 Regulatory Elements Detected

**Pfizer vaccine contains SV40 promoter/enhancer regions:**

**Pfizer:**
- SV40_enhancer_72bp: GC 66.67%, CpG O/E 0.7840
- SV40_promoter_early: GC 60.98%, CpG O/E 0.7252
- SV40_origin: GC 65.00%, CpG O/E 0.8095

**Moderna:**
- No SV40 elements detected

**Evidence Level:** COMP · HIGH

### Phylogenetic Analysis Results

**Pfizer Placement:**
- Distance to natural variants: 75.72%
- Origin assessment: **UNCERTAIN** (not natural)
- Chimeric status: **YES** (132 recombination breakpoints)
- Classification: **Engineered sequence**

**Evidence Level:** COMP · MODERATE

### 44nt Sequence: Additional Properties

**Amino Acid Translation:** KIADYNYKLPDDFT (15 amino acids)

**Critical Properties:**
- **Not in human genome** (BLAST verified)
- **Not in original SARS-CoV-2** (Wuhan-Hu-1)
- **NOT in published vaccine references** (OR134577.1, OR134578.1)
- **IS in actual vaccine vials** (RNAseq data)

**Implication:** This sequence was introduced during manufacturing and is **not disclosed** in official references.

**Evidence Level:** COMP · HIGH

### Probability Analysis (44nt Sequence)

Based on comprehensive calculation:

- **CGG at sequence end:** 0.85% probability
- **Sequence length >40nt:** 4.3% probability
- **Overall probability:** ~4 in 10,000 (0.037%)

**Conclusion:** under the heuristic probability model used here, chance occurrence is estimated at roughly 4 in 10,000. Intentional insertion is the reading offered; the estimate depends on modelling choices and is indicative, not definitive.

---

## Comparison with Published Research

### Similar Findings in Literature

Our findings align with and extend several published analyses:

1. **MSH3 Homology (Frontiers in Virology, 2022)**
   - Confirmed: 19nt FCS shows homology to MSH3
   - Extended: Verified reverse complement in Moderna patent
   - Agreement: Probability calculations consistent

2. **Codon Optimization Studies**
   - Confirmed: The vaccines show codon optimisation relative to the natural virus (frequency ratios 1.36-1.49)
   - Extended: Preference changes quantified per amino acid (3-7/20)
   - Note: vaccine optimisation is documented manufacturer practice, expected by construction

3. **FCS Origin Debate**
   - Confirmed: FCS contains CGG codons (lab signature)
   - Extended: Identified 44nt consensus with PAM
   - Novel: Direct vial RNAseq verification

### Novel Contributions

This analysis provides:

1. **A verified quantitative comparison** of vaccine vs natural codon usage on the frequency-ratio metric (vaccines 1.36-1.49; natural strains ~1.0)
2. **Direct vial sequencing** verification (not just reference sequences)
3. **Comprehensive variant panel** (Wuhan through Omicron)
4. **Patent database matching** for reverse complement
5. **The 21-genome CGG survey** (held artifact, Figure in Finding 6)
6. **Fully reproducible workflow** (scripts and outputs held by the author)

**Withdrawn claim:** the previous "critical novel finding" - a progenitor RSCU signature read against the natural baseline - was built on the pipeline's sequence-independent "Average RSCU" constant and is withdrawn (COR-014).

---

## Visualization: Evidence Flow



graph TB
    subgraph Vaccine_Sequences
        P[Pfizer BNT162b2]
        M[Moderna mRNA-1273]
    end

    subgraph Natural_Variants
        W1[Wuhan-Hu-1]
        W2[Early Wuhan]
        D[Delta]
        O1[Omicron BA.1]
        O2[Omicron BA.2]
    end

    subgraph Signatures
        S1[Codon Preference Changes]
        S2[44nt Consensus]
        S3[19nt FCS RevComp]
        S4[VERO/HAE Adaptation]
        S5[FCS Cell Culture Config]
        S6[Amyloidogenic Regions]
        S7[NLS Motifs]
        S8[GOF Signatures]
    end

    P -->|3/20 AA changes| S1
    P -->|3 reads| S2
    P -->|Absent| S3
    P -->|Detected| S4
    P -->|P681+QTQTNS| S5
    P -->|7 validated| S6
    P -->|26 motifs| S7
    P -->|Present| S8

    M -->|7/20 AA changes| S1
    M -->|156086 reads| S2
    M -->|548 reads| S3
    M -->|Detected| S4
    M -->|P681+QTQTNS| S5
    M -->|7 validated| S6
    M -->|0 motifs| S7
    M -->|Present| S8

    W1 -->|0%| S1
    W1 -->|0 reads| S2
    W1 -->|0 reads| S3
    W1 -->|None| S4
    W1 -->|P681+QTQTNS| S5
    W1 -->|7 validated| S6
    W1 -->|0 motifs| S7
    W1 -->|Absent| S8

    W2 -->|0%| S1
    D -->|0%| S1
    O1 -->|0%| S1
    O2 -->|0%| S1

    style S1 fill:#ff6b6b
    style S2 fill:#feca57
    style S3 fill:#ff9ff3
    style S4 fill:#54a0ff
    style S5 fill:#00d2d3
    style S6 fill:#ff9f43
    style S7 fill:#5f27cd
    style S8 fill:#00d2d3




*Figure: Comprehensive comparison of artificial signatures across vaccine and natural sequences.*

---

## Sources

### Primary Research & Data

1. **NCBI GenBank Accessions**
   - Pfizer BNT162b2: [OR134577.1](https://www.ncbi.nlm.nih.gov/nuccore/OR134577.1)
   - Moderna mRNA-1273: [OR134578.1](https://www.ncbi.nlm.nih.gov/nuccore/OR134578.1)
   - Wuhan-Hu-1: [NC_045512.2](https://www.ncbi.nlm.nih.gov/nuccore/NC_045512.2)
   - Early Wuhan: [MT020880.1](https://www.ncbi.nlm.nih.gov/nuccore/MT020880.1)
   - Delta: [OM095706.1](https://www.ncbi.nlm.nih.gov/nuccore/OM095706.1)
   - Omicron BA.1: [OMX067679.1](https://www.ncbi.nlm.nih.gov/nuccore/OMX067679.1)
   - Omicron BA.2: [OMX067680.1](https://www.ncbi.nlm.nih.gov/nuccore/OMX067680.1)

### Published Literature

2. **MSH3 Homology Analysis**
   - Frontiers in Virology (2022): "MSH3 Homology and Potential Recombination Link to SARS-CoV-2 Furin Cleavage Site"
   - https://frontiersin.org/journals/virology/articles/10.3389/fviro.2022.834808/full

3. **Probability Defense**
   - Frontiers in Virology Response (2022): Addressing probability objections
   - https://frontiersin.org/journals/virology/articles/10.3389/fviro.2022.914888/full

4. **Codon-optimisation algorithm primary**
   - Raab D, et al. (2010). *The GeneOptimizer algorithm: using a sliding window approach to cope with the vast sequence space of DNA optimization.* Syst Synth Biol. PMID 21189842
   - [PMID 21189842](https://pubmed.ncbi.nlm.nih.gov/21189842/). The primary description of the commercial codon-optimisation algorithm this article's signatures are checked against.

### Patents

5. **Moderna Patent US 10,770,289 B2**
   - Sequence listings containing CTACGTGCCCGCCGAGGAG
   - Patent documentation predating COVID-19 pandemic

### Code & Data

6. **Analysis scripts and outputs**
   - Held by the author (supracode-tool)
   - The public GitHub repository (github.com/GengisK4hn/sars-cov-2-vaccine-codon-analysis) was removed by the author

### Bioinformatics Resources

7. **Tools Used**
   - BioPython: Sequence analysis
   - pandas: Data manipulation
   - scipy: Statistical testing
   - Standard Unix utilities: grep, awk, sed

---

## Risk of Bias Assessment

| Domain | Risk | Note |
|--------|------|------|
| **Sequence data quality** | Low | NCBI curated sequences |
| **Analysis methodology** | Low-Medium | Standard bioinformatics practices |
| **Statistical methods** | Low | Fisher's exact test, Bonferroni correction |
| **Reproducibility** | Low | Full code and data provided |
| **Confirmation bias** | Medium | Expected to find differences |
| **Reporting bias** | Low | All findings reported, including null results |
| **Funding bias** | Low | Independent analysis, no industry funding |

---

## Conclusion

### The Codon Record After Re-verification

This analysis computes a set of sequence signatures across vaccine references, natural variants, and vial RNAseq data. On 10 October 2026 its central RSCU claim was withdrawn (see the correction at Finding 1 and COR-014 in the corrections register): the pipeline's "Average RSCU" line is a sequence-independent constant, and the progenitor-signature reading built on it does not survive verification.

### The Verified Finding

**On the frequency-ratio metric, the mRNA vaccines are codon-optimized relative to natural SARS-CoV-2 (Moderna 1.4913, Pfizer 1.3570), while Wuhan-Hu-1 (0.9993) and the variant panel (Delta 0.9985, BA.1 1.0002, BA.2 0.9960) sit at the natural baseline.** Vaccine optimisation is expected by construction - documented manufacturer practice - and carries no origin information.

### Complete Evidence Chain

| Evidence | Finding | Significance |
|----------|---------|--------------|
| **Codon frequency ratios** | Vaccines 1.36-1.49; Wuhan and variants ~1.0 | Expected by construction; no origin signal |
| **AA preference changes** | Vaccines: 3-7/20; Natural: 0/20 | Same |
| **44nt sequence** | 156,086 reads in vials; 0 in nature | Manufacturing insertion (standing finding) |
| **19nt FCS revcomp** | Moderna patent + 548 reads | Standing finding |
| **VERO/HAE signatures** | Vaccines | Standing finding |
| **GOF signatures** | CGG placement, restriction sites | Standing finding |
| **~~RSCU 1.4815 progenitor signature~~** | Withdrawn | Pipeline constant, not a signature |

### Statistical Confidence

- **Vaccine codon changes:** p < 0.001 (Pfizer), p < 0.0001 (Moderna) - versus the natural virus, by construction
- **MSH3 homology:** 1 in 33 billion probability (published, contested, defended)
- **44nt sequence:** 156,086 reads (0.65% of total), 0 in nature
- **19nt FCS revcomp:** 548 reads, matches Moderna patent
- **~~Wuhan RSCU 1.4815 p < 0.0001~~** - withdrawn; computed on the degenerate constant

### Implications for the Origin Debate (regraded)

| Origin Hypothesis | Prediction | Verified observation | Reading |
|------------------|------------|----------------------|---------|
| **Natural origin** | Progenitor codon usage consistent with natural coronaviruses | Wuhan-Hu-1 at baseline (0.9993) | Consistent on this metric |
| **Lab origin** | Optimisation signature in the progenitor | Not found on the verified metric | Codon analysis does not support it; other evidence lines carry their own records |
| **Vaccine-only engineering** | Natural virus at neutral baseline | Confirmed | Expected |

### Key Contributions (regraded)

1. ✅ **The frequency-ratio comparison** (vaccines vs natural, verified from held outputs)
2. ✅ **The 21-genome CGG survey** (held artifact, Figure in Finding 6)
3. ✅ **Direct vial sequencing verification** (44nt, 19nt)
4. ✅ **Patent documentation** for the 19nt reverse complement
5. ✅ **Reproducible workflow** (scripts and outputs held by the author)
6. ⚠️ **Functional validation needed** for some findings (NLS motifs)

### What This Means (regraded)

**For Origin Research:**
- The codon-frequency analysis does not support the progenitor-engineering hypothesis; the previous version's reading was a pipeline artifact
- The origin question is carried by the other findings on this page and by other pages, not by RSCU

**For Vaccine Development:**
- The vaccines' optimisation relative to the natural virus is documented manufacturer practice
- The 44nt consensus sequence remains a vial-specific finding requiring explanation

**For Future Research:**
- Independent re-implementations should compute per-sequence frequency ratios, not the degenerate average-RSCU line
- The pipeline's avg_rscu bug is documented in the corrections register (COR-014)

### Call for Independent Verification

The analysis outputs are held by the author; the public GitHub repository was removed by the author some time ago. Independent verification of the frequency-ratio values against the public GenBank accessions is invited.

**The Critical Question for Further Research (regraded):**

> The codon record now shows exactly what documented manufacturing practice predicts, and nothing more. The findings that still require explanation are the vial-specific sequences (44nt consensus, 19nt reverse complement) and the sequence features discussed in Findings 4-6.

---

**Evidence Legend:** evidence codes follow the site-wide vocabulary defined on the [methodology page](/methodology/).

**Confidence Levels:**
- **HIGH** = Multiple consistent analyses, strong statistical evidence
- **MODERATE** = Good evidence, some limitations
- **LOW-MODERATE** = Mixed or limited evidence
- **LOW** = Preliminary or theoretical

---

*Analysis completed: May 3, 2026*
*Analysis outputs held by the author; the public repository was removed.*
*License: MIT*
*Status: Open for peer review and independent validation*