---

name: variant-interpretation-acmg description: Classifies genetic variants according to ACMG (American College of Medical Genetics) guidelines. license: MIT metadata: author: AI Group version: "1.0.0" compatibility:

• system: Python 3.10+ allowed-tools:
• run_shell_command
• read_file

---

Variant Interpretation (ACMG)

The Variant Interpretation Skill automates the classification of genetic variants (Pathogenic, Benign, VUS) using a rules-based engine derived from ACMG guidelines.

When to Use This Skill

• When analyzing a VCF file for clinical reporting.
• To determine the clinical significance of a specific mutation (e.g., BRCA1 c.123A>G).
• To aggregate evidence (population freq, computational predictions) into a final verdict.

Core Capabilities

1. Rule Scoring: Applies codes like PVS1 (Null variant), PM2 (Rare), PP3 (In silico).
2. Classification: Combines scores to reach a verdict (Pathogenic, Likely Pathogenic, VUS, etc.).
3. Explanation: Provides the logic/evidence used for the classification.

Workflow

1. Input: Variant details (Gene, HGVS, Consequence) or Evidence codes directly.
2. Process: Sums weights of applied ACMG criteria.
3. Output: Final classification and score breakdown.

Example Usage

User: "Classify a variant with evidence PVS1 and PM2."

Agent Action:

Copypython3 Skills/Genomics/Variant_Interpretation/acmg_classifier.py \
    --evidence "PVS1,PM2"