claw-methylation-cycle

Files (5)

• SKILL.md
• demo_input.txt
• methylation_cycle.py
• test_methylation_cycle.py · tests/
• test_reproducibility_bundle.py · tests/

SKILL.md (first lines)

# claw-methylation-cycle

Methylation cycle analysis skill for ClawBio. Produces enzymatic activity
profiles, Net Methylation Capacity (NMC), BH4 axis estimates, compound
heterozygosity detection, and clinician-review genotype findings from raw
SNP genotype data.

---

## Trigger

**Fire this skill when:**

- The user asks about methylation, MTHFR variants, folate cycle, or
  homocysteine risk from a genotype file.
- A raw SNP file (23andMe / ADNTRO / Ancestry format) is provided and the
  clinical question involves methylation, BH4, dopamine/serotonin synthesis
  capacity, or neurodevelopmental contexts (ADHD, depression, anxiety).
- The upstream workflow (PharmGx Reporter, NutriGx Advisor) has flagged
  MTHFR or MTRR and the clinician needs the full methylation panel.
- Keywords present: `methylation`, `MTHFR`, `BH4`, `folate cycle`,
  `metilación`, `ciclo de metilación`, `homocysteine`, `5-MTHF`,
  `methylcobalamin`, `neurotransmitter synthesis`, `dopamine upstream`.

**Do NOT fire this skill when:**

- The question is purely about folic acid supplementation without a genotype file.
- The user is asking about MTHFR in the context of thrombophilia/clotting
  only — use PharmGx Reporter for warfarin/anticoagulation questions.
- Only N-GENE polygenic risk data is available (no raw SNP file) — this skill
  requires genotype-level input; PRS percentiles are not sufficient.
- The SNP file format is VCF, FASTQ, BAM, or PLINK binary — these require
  preprocessing before this skill can run.
- The clinical question is exclusively pharmacogenomic (CYP enzymes) — use
  PharmGx Reporter instead.

---

## Workflow

1. **Receive input** — Accept either a raw genotype file path or a pre-parsed
   `snp_dict`. If a file is provided, call `parse_genotype_file()` to extract
   the rsID → genotype mapping.

2. **Panel coverage check** — Compare detected rsIDs against the 9-gene
   methylation panel. Log missing SNPs. For any SNP absent from the input,
   mark the corresponding gene as `not_assessed` — do NOT silently assume
   normal activity (Safety Rule 6).

3. **Enzymatic activity scoring** — For each gene, map the diplotype to an
   estimated activity percentage. Heterozygous risk variants reduce activity
   by their assigned weight; homozygous variants apply the full reduction.

4. **Compound heterozygosity detection** — Check MTHFR C677T (rs1801133) and
   A1298C (rs1801131) simultaneously. If both are heterozygous, set
   `compound_heterozygosity = True` and apply the combined activity reduction
   (~15% of normal — more severe than either variant alone).

5. **Net Methylation Capacity (NMC)** — Compute the weighted average of all
   enzyme activities. Clamp to [0, 100]. Expose `coverage_pct` and
   `snps_missing`; flag NMC as partial if key SNPs are absent.

6. **BH4 axis capacity** — Derive BH4 from MTHFR activity and MTRR modifier.
   Report clinical implications for dopamine and serotonin synthesis in
   neurodevelopmental presentations.

7. **Prioritised recommendations** — Generate PRIORITY 1 / 2 / 3
   recommendations based on active findings. Lead with highest clinical
   impact (compound het MTHFR or severely reduced BH4).

8. **Output** — Write `report.md` (human-readable) and `result.json`
   (structured, for downstream integration).

---

## Example Output

```
╔══════════════════════════════════════════════════════════════╗
║  ClawBio · Methylation Cycle Analysis Report                 ║
║  Author: Samuel Carmona Aguirre · RUO — Not a medical device ║
╚══════════════════════════════════════════════════════════════╝

Executive Summary
─────────────────
Net Methylation Capacity : 53 / 100  🔴 Reduced
BH4 Axis Capacity        : 31 / 100  🔴 Reduced
MTHFR Compound Het.      : YES ⚠️   (C677T + A1298C)
Dopamine Synthesis       : Severely Reduced
Serotonin Synthesis      : Severely Reduced

Enzymatic Activity Profile
──────────────────────────
Gene    Activity  Status                Key Variants
MTHFR    15%     🔴 Severely reduced   C677T, A1298C
MTRR     60%     🟡 Moderately reduced A66G
MTR     100%     🟢 Normal             –
CBS     100%     🟢 Normal             –
BHMT     40%     🔴 Moderately reduced R239Q
SHMT1    80%     🟢 Mildly reduced     C1420T
COMT     55%     🟡 Moderately reduced Val158Met
AHCY    100%     🟢 Normal             –

Clinical Recommendations — FOR CLINICIAN REVIEW ONLY
──────────────────────────────────────────────────────
⚠️  The following is genotype-based information for qualified clinician use.
    Do not self-administer. All nutrients listed are reported in the peer-reviewed
    literature for the pathways indicated; dosing and indication require
    individualised clinical assessment.

Genotype findings:
  • 5-MTHF (methylfolate) preferred over synthetic folic acid (MTHFR C677T/A1298C).
    Ref: Lamers Y et al. (2004) Am J Clin Nutr 80(5):1234-41.
  • MTHFR compound het: methylcobalamin co-administration reported in literature.
    Ref: Ledford AW et al. (2021) Nutrients 13(3):768.
  • BH4 capacity at 31%: riboflavin (B2) reported as MTHFR cofactor supporting BH4
    regeneration. Ref: McNulty H et al. (2017) Am J Clin Nutr 106(1):128-36.
  • MTRR A66G: methylcobalamin preferred over cyanocobalamin per functional studies.
    Ref: Olteanu H et al. (2002) Biochemistry 41(45):13378-85.

From ClawBio/ClawBio, MIT.

More from ClawBio/ClawBio