Look up clinical significance and associated diseases for a genetic variant in ClinVar, given its gene symbol and short-form protein change (e.g. "BRAF", "V600E"). Call get_vep_consequence first and pass its gene_symbol and protein_change_short values as arguments here — not protein_change (the p.Val600Glu-style HGVS format), which will not match. Returns a dict with: - found: True if results are found or False - matches: List filled with dict of classifications and traits: - "variation_id": Clinvar variation ID - "title": Variant title (e.g. 'NM_004333.6(BRAF):c.1799T>A (p.Val600Glu)') - "germline_classification": (e.g. 'Conflicting classifications of pathogenicity') - "germline_traits": (e.g. ['Cardiovascular phenotype', 'Vascular malformation', 'RASopathy']) - "clinical_impact_classification": How clinically impactful (e.g. 'Tier I - Strong') - "clinical_impact_traits": (e.g. ['Ganglioglioma', 'Malignant glioma', 'Dysembryoplastic neuroepithelial tumor']) - "oncogenicity_classification": Tendency to cause tumors/cancer (e.g. 'Oncogenic') - "oncogenicity_traits": (e.g. ['Neoplasm']) If found is False, function returns {"found": False, "matches": []} Each match includes "position_verified": true if the genomic position matches what get_vep_consequence reported, false if it doesn't (indicating possible ambiguity or a different variant), or null if verification wasn't possible. Treat false results with caution and mention the discrepancy to the user.

Look up the biological effect of a genetic variant given in supported HGVS notation. Use genomic HGVS (e.g. "chr7:g.140753336A>T") or coding HGVS with a transcript accession (e.g. "NM_004333.6:c.1799T>A"). Bare coding HGVS like "c.1799T>A" is ambiguous and unsupported. Use this when a user asks what a specific DNA variant does. Genome build and transcript scope (MANE Select only vs. all transcripts) are set by the user via CLI commands, not by this tool's arguments. Returns a dict with: - gene_symbol: the affected gene (e.g. "BRAF") - consequence: the molecular consequence type (e.g. "missense_variant") - protein_change: the amino acid change in HGVS protein notation (e.g. "p.Val600Glu") ("None" if not applicable, like for synonymous variants) - protein_change_short: the same change in short form, (e.g. "V600E" — use this exact value when calling get_clinvar_summary) - impact: Ensembl's severity bucket for the consequence type (HIGH/MODERATE/LOW/MODIFIER) - sift_prediction / sift_score: SIFT's tolerance prediction and probability (lower score = more likely damaging; "None" if not applicable to this consequence type) - polyphen_prediction / polyphen_score: PolyPhen-2's damage prediction and probability (higher score = more likely damaging; "None" if not applicable) - build: genome build used for this lookup ("grch38" or "grch37") - input_hgvs: the original HGVS input string - hgvs_format: "genomic" for g. inputs or "coding" for transcript-qualified c. inputs - transcripts: only present if the user has selected "all transcripts" mode and the variant has more than one transcript consequence. A list of dicts, each with transcript_id, consequence, protein_change, impact, sift_prediction, and polyphen_prediction for that transcript. The top-level fields above always reflect the primary (MANE Select or first-returned) transcript, which is also what gets passed to get_clinvar_summary — do not use a value from the transcripts list for that purpose.