COMPUTE, DON'T DESCRIBE

When analysis requires computation (statistics, data processing, scoring, enrichment), write and run Python code via Bash. Don't describe what you would do — execute it and report actual results. Use ToolUniverse tools to retrieve data, then Python (pandas, scipy, statsmodels, matplotlib) to analyze it.

Structural Variant Analysis Workflow

Systematic analysis of structural variants (deletions, duplications, inversions, translocations, complex rearrangements) for clinical genomics interpretation using ACMG-adapted criteria.

LOOK UP DON'T GUESS - Always retrieve ClinGen HI/TS scores, gnomAD frequencies, and ClinVar evidence from tools. Do not infer dosage sensitivity from gene function alone.

KEY PRINCIPLES:

1. Report-first approach - Create SV_analysis_report.md FIRST, then populate progressively
2. ACMG-style classification - Pathogenic/Likely Pathogenic/VUS/Likely Benign/Benign with explicit evidence
3. Evidence grading - Grade all findings by confidence level (High/Moderate/Limited)
4. Dosage sensitivity critical - Gene dosage effects drive SV pathogenicity
5. Breakpoint precision matters - Exact gene disruption vs dosage-only effects
6. Population context essential - gnomAD SVs for frequency assessment
7. English-first queries - Always use English terms in tool calls. Respond in the user's language