Clinical Variant Interpreter

Systematic variant interpretation using ToolUniverse - from raw variant calls to ACMG-classified clinical recommendations with structural impact analysis.

Triggers

Use this skill when users:

• Ask about variant interpretation, classification, or pathogenicity
• Have VCF data needing clinical annotation
• Need ACMG classification for variants
• Want structural impact analysis for missense variants

Key Principles

1. ACMG-Guided - Follow ACMG/AMP 2015 guidelines with explicit evidence codes
2. Structural Evidence - Use AlphaFold2 for novel structural impact analysis
3. Population Context - gnomAD frequencies with ancestry-specific data
4. Actionable Output - Clear recommendations, not just classifications
5. English-first queries - Always use English terms in tool calls; respond in user's language

---

LOOK UP, DON'T GUESS

When asked about a variant's significance, query ClinVar/gnomAD/CIViC FIRST. Never classify a variant without checking databases. When you're not sure about a fact, your first instinct should be to SEARCH for it using tools, not to reason harder from memory.

---

Workflow Overview

Phase 1: VARIANT IDENTITY        → Normalize HGVS, map gene/transcript/consequence
Phase 2: CLINICAL DATABASES       → ClinVar, gnomAD, OMIM, ClinGen, COSMIC, SpliceAI
Phase 2.5: REGULATORY CONTEXT     → ChIPAtlas/ENCODE annotation + DL variant-effect (AlphaGenome/Enformer/Borzoi/ChromBPNet/Evo2) (non-coding only)
Phase 3: COMPUTATIONAL PREDICTIONS → CADD, AlphaMissense, EVE, SIFT/PolyPhen
Phase 4: STRUCTURAL ANALYSIS      → PDB/AlphaFold2, domains, functional sites (VUS/novel)
Phase 4.5: EXPRESSION CONTEXT     → CELLxGENE, GTEx tissue expression
Phase 5: LITERATURE EVIDENCE      → PubMed, EuropePMC, BioRxiv, MedRxiv
Phase 6: ACMG CLASSIFICATION      → Evidence codes, classification, recommendations

---

Phase 1: Variant Identity

Tools: MyVariant_query_variants, EnsemblVar_get_variant_consequences, NCBIGene_search, VariantValidator_gene2transcripts, VariantValidator_validate_variant, Tark_get_mane_transcripts, Tark_get_transcript

VariantValidator_gene2transcripts: Look up MANE Select and MANE Plus Clinical transcripts for a gene. Use this to identify the correct canonical transcript before variant annotation.

• Parameters: gene_symbol (e.g. "TP53"), transcript_set ("mane" | "refseq" | "ensembl" | "all"), genome_build ("GRCh38" default)
• Returns: Array of {current_symbol, transcripts: [{reference, annotations: {mane_select, mane_plus_clinical}}]}
• Aliases: gene and gene_name also accepted for gene_symbol

Tark_get_mane_transcripts: Lightweight MANE Select / MANE Plus Clinical lookup from Ensembl Tark with the ENST↔RefSeq pairing (e.g. gene="BRCA2" → ENST00000380152.8 / NM_000059.4). Use as a quick cross-check of the canonical transcript namespace alongside VariantValidator, or to translate an ENST↔NM accession. Tark_get_transcript (param stable_id, e.g. "ENST00000380152") returns the archived transcript record (assembly, biotype, coordinates, per-release versions) when you need to resolve a specific transcript version.

VariantValidator_validate_variant: Validate HGVS variant descriptions and get normalized notation with genomic/transcript/protein consequences.

• Parameters: genome_build ("GRCh37" | "GRCh38"), variant_description (HGVS, e.g. "NM_007294.4:c.5266dup"), select_transcripts (transcript or "all")
• Returns: Validated HGVS, protein consequence, genomic coordinates, gene IDs

Capture: HGVS notation (c. and p.), gene symbol, canonical transcript (MANE Select via VariantValidator), consequence type, amino acid change, exon/intron location.

Phase 2: Clinical Databases

Tools: ClinVar_search_variants, gnomad_search_variants, gnomad_get_variant, OMIM_search, OMIM_get_entry, ClinGen_search_gene_validity, ClinGen_search_dosage_sensitivity, ClinGen_search_actionability, COSMIC_search_mutations, COSMIC_get_mutations_by_gene, DisGeNET_search_gene, DisGeNET_get_vda, SpliceAI_predict_splice, SpliceAI_get_max_delta, civic_get_variants_by_gene, civic_search_evidence_items, civic_search_assertions

gnomAD two-step workflow: gnomad_search_variants only accepts rsIDs or variant IDs (not gene names). Search by rsID first, then use the returned variant_id with gnomad_get_variant to get population allele frequencies.

CIViC: Use civic_search_genes(query="<gene_symbol>") to find the CIViC gene ID dynamically (do NOT rely on a hardcoded lookup table). Then use civic_get_variants_by_gene(gene_id=<id>) and civic_search_evidence_items for actionability details. If civic_search_genes returns no results, the gene may not be curated in CIViC — note this gap.

OncoKB note: Demo mode only supports BRAF, TP53, ROS1. For other genes, set ONCOKB_API_TOKEN environment variable.

Use SpliceAI for: intronic variants near splice sites, synonymous variants, exonic variants near splice junctions.

See CODE_PATTERNS.md for implementation details.

Phase 2.5: Regulatory Context (Non-Coding Only)

Apply for intronic (non-splice), promoter, UTR, or intergenic variants near disease genes.

Annotation — what regulatory element is here: ChIPAtlas_enrichment_analysis, ChIPAtlas_get_peak_data, ENCODE_search_experiments, ENCODE_get_experiment. These tell you whether the variant falls in a known TF-binding peak, enhancer, or open-chromatin region.

Prediction — what the variant does to regulation: annotation says an element is present, not whether this specific allele disrupts it. Sequence-based deep-learning models answer that directly: they read the reference and alternate DNA windows and predict the change in regulatory signal. This is what turns "the variant is in an enhancer" into "the variant is predicted to reduce accessibility/expression in the relevant tissue" — the mechanistic evidence ACMG PS3/PP3 actually needs for a non-coding variant, where SIFT/PolyPhen/AlphaMissense do not apply.

Tool	Predicts	Context	Access
AlphaGenome_score_variant	Δ across RNA-seq / ATAC / CAGE / splice tracks (frontier accuracy, single-base)	up to 1 Mb	hosted API — needs ALPHA_GENOME_API_KEY
run_enformer_variant_effect	Δ across 5,313 human tracks (expression, chromatin, TF binding)	196 kb	remote MCP server
run_borzoi_variant_effect	Δ in RNA-seq coverage (expression / polyA / splicing emphasis)	524 kb	remote MCP server
run_chrombpnet_variant_effect	Δ in chromatin accessibility (ATAC / DNase), base-resolution	~2 kb	remote MCP server
Evo2_score_variant	Genome-foundation-model delta log-likelihood; covers coding and non-coding	up to 1 Mb	hosted NIM — needs NVIDIA_API_KEY

Reading the score: these return Δ (alt − ref) effect sizes, not calibrated pathogenicity probabilities. A large predicted disruption in a tissue-relevant track is mechanistic support (PS3_supporting / PP3) for a non-coding variant; near-zero across tracks supports BP4. Rank or calibrate against known regulatory variants rather than applying an absolute cutoff.

Which to pick: start with AlphaGenome_score_variant (broadest readout, longest context, frontier accuracy) when its key is set; run_enformer_variant_effect / run_borzoi_variant_effect are the named, self-hostable equivalents (Enformer for general regulation, Borzoi when expression/splicing is the question); run_chrombpnet_variant_effect when the hypothesis is specifically chromatin accessibility; Evo2_score_variant as a sequence-only check that also works on coding variants. If no key/server is provisioned, fall back to the ChIPAtlas/ENCODE annotation above and note the predictive gap rather than guessing.

Inputs: AlphaGenome_score_variant takes chromosome + position + reference_bases/alternate_bases (+ output_type, sequence_length); Evo2_score_variant takes a DNA window as sequence + position + alternate (point substitution) or ref_sequence/alt_sequence, plus optional model (evo2-40b default, evo2-7b faster); the Enformer/Borzoi/ChromBPNet remote tools take the variant locus and score the change over their output tracks.

Phase 2.9: Short-Circuit Check

Before full ACMG classification, check if the variant already has an expert panel classification in ClinVar. Use MyVariant_query_variants with the rsID or HGVS notation — the clinvar field in the response includes clinical significance, review status, and RCV records. If an expert panel has already classified the variant as Pathogenic or Benign, note this prominently and focus on confirming/contextualizing rather than de novo classification.

Phase 3: Computational Predictions

Primary approach: MyVariant_query_variants with fields=dbnsfp,clinvar,cadd,gnomad_genome retrieves 15+ predictor scores (SIFT, PolyPhen, CADD, REVEL, AlphaMissense, MetaRNN, FATHMM, GERP, PhyloP, etc.) in a single call. This is usually sufficient.

REVEL/AlphaMissense fallback: If MyVariant_query_variants returns no dbnsfp block, use the dedicated tool:

1. MyVariant_get_pathogenicity_scores (PREFERRED FALLBACK) — returns REVEL, AlphaMissense, SIFT, PolyPhen2, MetaRNN, GERP, PhyloP, and more in a single call with pre-configured dbnsfp fields. Input: variant_id (rsID or HGVS genomic).
2. CADD_get_variant_score (PHRED 0-99) — works for most variants
3. AlphaMissense_get_variant_score (0-1, needs UniProt ID) — missense only
4. EVE_get_variant_score (0-1) — missense only
5. EnsemblVEP_annotate_hgvs (VEP with colocated variants) — includes SIFT/PolyPhen
6. If REVEL is still unavailable, note this as a limitation and rely on CADD + SIFT + PolyPhen consensus. REVEL absence does not prevent classification.

Consensus: Run CADD (all variants) + AlphaMissense + EVE (missense). 2+ concordant damaging = strong PP3; 2+ concordant benign = strong BP4.

See ACMG_CLASSIFICATION.md for thresholds.

Phase 4: Structural Analysis (VUS/Novel Missense)

Tools: PDBe_get_uniprot_mappings, NvidiaNIM_alphafold2 (requires NVIDIA_API_KEY env var; free key at build.nvidia.com), alphafold_get_prediction (param: qualifier, e.g., UniProt accession), InterPro_get_protein_domains, UniProt_get_function_by_accession

Workflow: Get structure -> map residue -> assess domain/functional site -> predict destabilization.

AlphaFold size limitation: Very large proteins (>2,700 aa, e.g., BRCA2 at 3,418 aa) may not have AlphaFold predictions via the standard API. Fall back to published structural studies or PDBe_get_uniprot_mappings for experimental structures.

Phase 4.2: Mechanism of Effect (VUS missense, ESMC-6B SAE)

AlphaMissense / REVEL / CADD give a pathogenicity score but no mechanism. When you need to answer "how does this variant disrupt protein function" — e.g. for VUS write-ups, clinical reports, or to triangulate a discordant predictor consensus — use the ESMC-6B Sparse Autoencoder to identify which interpretable protein-language-model features the mutation disrupts.

One-call mechanism summary (recommended starting point):

mech = tu.tools.ESM_explain_variant_mechanism(
    sequence=wt_aa_sequence,   # full reference protein sequence
    position=600,              # 1-indexed
    ref_aa="V",
    alt_aa="E",
    top_k_features=5,          # describe top 5 lost + top 5 gained
)
# mech["data"]["mechanism_summary"] e.g.:
#   "Disrupted feature categories (lost): catalytic=2, ligand-binding=1;
#    Induced feature categories (gained): structural-stability=1"

Returns mechanism_summary, per-feature lost/gained tables, and category aggregates. Use the category aggregate to support or qualify the pathogenicity verdict in the report:

• catalytic / ligand-binding / ptm lost → mechanistic support for PP3
• secondary-structure / structural-stability gained on a stable WT region → mechanistic basis for "destabilizing" claim
• No interpretable change at top-K → does not weaken AlphaMissense alone, but flag for caution

When you have a saturation question (e.g. "score all 19 substitutions at residue 600 to find the most disruptive"): use ESM_score_variant_sae_batch — 1 Forge call for the reference + 1 per variant, instead of 2 per variant.

When the region is what matters (e.g. "what's the SAE signature of the kinase activation loop, residues 754-771"): use ESM_get_region_sae_features then ESM_describe_sae_feature on the top hits.

Requires: ESM_API_KEY env var (free non-commercial token at https://forge.evolutionaryscale.ai) and pip install 'esm @ git+https://github.com/evolutionaryscale/esm@ee891c52' (SAE support is on an unmerged feature branch — PyPI esm 3.2.x does NOT include SAEConfig). License: EvolutionaryScale Cambrian Inference License — non-commercial use only.

Phase 4.5: Expression Context

Tools: CELLxGENE_get_expression_data, CELLxGENE_get_cell_metadata, GTEx_get_median_gene_expression

Confirms gene expression in disease-relevant tissues. Supports PP4 if highly restricted; challenges classification if not expressed in affected tissue.

Phase 5: Literature Evidence

Tools: PubMed_search_articles, EuropePMC_search_articles, BioRxiv_list_recent_preprints, MedRxiv_get_preprint, openalex_search_works, SemanticScholar_search_papers

Always flag preprints as NOT peer-reviewed.

Phase 6: ACMG Classification

Apply all relevant evidence codes (PVS1, PS1, PS3, PM1, PM2, PM5, PP3, PP5 for pathogenic; BA1, BS1, BS3, BP4, BP7 for benign). See ACMG_CLASSIFICATION.md for the complete algorithm.

Gene-Specific Population Frequency Thresholds

BS1 (allele frequency too high for disorder) requires gene-specific calibration, not a universal cutoff:

• High-penetrance genes (BRCA1, TP53): BS1 threshold ~0.0001
• Moderate-penetrance genes (PALB2, ATM, CHEK2): BS1 threshold ~0.001
• Low-penetrance/common disease genes: BS1 threshold higher, depends on disease prevalence
• Formula: BS1 threshold = (disease prevalence × max allelic contribution × max genetic contribution) / penetrance
• When in doubt, compare the variant's AF to the highest AF of any known pathogenic variant in the same gene — if it exceeds that, BS1 is likely applicable.

Handling Conflicting Evidence: Functional vs Epidemiological

This is one of the most challenging scenarios in variant interpretation. When a biochemical assay shows damage but population/epidemiological data shows no disease association:

1. Epidemiological data generally trumps in-vitro assays for clinical classification. A variant found at ~0.1% frequency with no disease association in 40K+ cases is unlikely to be clinically significant, even if it reduces protein function in a tube.
2. Apply PS3/BS3 carefully: ClinGen's SVI recommends that PS3 (functional evidence for pathogenicity) requires the assay to be validated against known pathogenic AND known benign controls. A single biochemical study without such validation is PS3_Supporting at best.
3. Hypomorphic variants: Some variants genuinely reduce protein function (detectable in sensitive assays) but not enough to cause disease. This is biologically real and does not make them pathogenic.
4. Document the conflict explicitly in the report. State: "Biochemical assay X shows [result], but case-control study Y with N cases found no significant disease association. Per ACMG guidelines, the epidemiological evidence is weighted more heavily for clinical classification."

Bayesian ACMG Point System (Tavtigian et al. 2018)

Modern clinical labs use a point-based system instead of the original rule-counting approach:

Evidence Level	Pathogenic Points	Benign Points
Very Strong (PVS1)	+8	--
Strong (PS1-PS4)	+4 each	-4 each (BS1-BS4)
Moderate (PM1-PM6)	+2 each	--
Supporting (PP1-PP5)	+1 each	-1 each (BP1-BP7)
Stand-alone (BA1)	--	-8

Classification by total points:

• Pathogenic: >= 10 points
• Likely Pathogenic: 6-9 points
• VUS: -5 to 5 points
• Likely Benign: -6 to -9 points
• Benign: <= -10 points

This system handles conflicting evidence naturally — a variant with PS3 (+4) and BS1 (-4) and BP4 (-1) nets -1, which is VUS. The original rule-based approach struggles with this scenario.

Computational procedure: ACMG Bayesian classification

# Automated ACMG point calculation
# Input: dict of evidence codes with their applied strength

def classify_acmg(evidence: dict) -> dict:
    """
    Classify a variant using the Bayesian ACMG point system.

    Args:
        evidence: dict mapping ACMG codes to strength levels.
            Pathogenic codes: 'very_strong', 'strong', 'moderate', 'supporting'
            Benign codes: 'stand_alone', 'strong', 'supporting'

    Example:
        evidence = {
            'BS1': 'strong',       # AF too high
            'BS3': 'supporting',   # Epidemiological evidence against pathogenicity
            'BP6': 'supporting',   # ClinVar benign consensus
            'PP3': 'supporting',   # Computational predictors say damaging
        }
    """
    pathogenic_points = {
        'very_strong': 8, 'strong': 4, 'moderate': 2, 'supporting': 1
    }
    benign_points = {
        'stand_alone': -8, 'strong': -4, 'supporting': -1
    }

    total = 0
    details = []
    for code, strength in evidence.items():
        if code.startswith(('PVS', 'PS', 'PM', 'PP')):
            pts = pathogenic_points.get(strength, 0)
        elif code.startswith(('BA', 'BS', 'BP')):
            pts = benign_points.get(strength, 0)
        else:
            pts = 0
        total += pts
        details.append(f"{code} ({strength}): {pts:+d}")

    if total >= 10:
        classification = "Pathogenic"
    elif 6 <= total <= 9:
        classification = "Likely Pathogenic"
    elif -5 <= total <= 5:
        classification = "VUS"
    elif -9 <= total <= -6:
        classification = "Likely Benign"
    else:
        classification = "Benign"

    return {
        'classification': classification,
        'total_points': total,
        'evidence_breakdown': details
    }

# Example: PALB2 c.2816T>G (from test case)
result = classify_acmg({
    'BS1': 'strong',       # gnomAD AF 0.00105 exceeds threshold
    'BS3': 'supporting',   # Case-control study shows no association
    'BP6': 'supporting',   # ClinVar 13 submitters say benign/likely benign
})
# Output: Likely Benign, total_points=-6, evidence: BS1(strong):-4, BS3(supporting):-1, BP6(supporting):-1

Use this procedure after collecting all evidence from Phases 1-5 to compute the final classification.

Gene-Specific VCEP Criteria

ClinGen Variant Curation Expert Panels (VCEPs) publish gene-specific ACMG modifications. Before classifying, check if a VCEP exists:

• ClinGen_search_gene_validity(gene="<gene_symbol>") — if validity is "Definitive" or "Strong", a VCEP likely exists
• Common VCEPs: BRCA1/2 (Enigma), TP53, PTEN, CDH1, PALB2, RASopathies, Lynch syndrome genes
• VCEP criteria override generic ACMG criteria (e.g., PALB2 VCEP has specific PM1 hotspot regions)

Predictor Weighting

Not all computational predictors are equal. For missense variants:

• REVEL (AUC ~0.95) — best single meta-predictor; weight highest
• AlphaMissense (AUC ~0.94) — strong, structure-aware
• CADD (AUC ~0.85) — good for all variant types, but less specific for missense
• SIFT/PolyPhen (AUC ~0.80) — legacy tools; useful for consensus but not individually decisive

When predictors disagree: if REVEL says tolerated but SIFT/PolyPhen say damaging, lean toward REVEL. If REVEL is unavailable, require 3+ concordant predictions for PP3/BP4.

Tool Failure Fallbacks

If a primary tool fails, use these alternatives:

• ClinVar_search_variants returns 0 results: Use MyVariant_query_variants with rsID or HGVS — the clinvar field in MyVariant is more reliable for variant lookup than NCBI Entrez search
• gnomad_search_variants fails: Use EnsemblVEP_annotate_hgvs which includes gnomAD frequency via colocated variants
• CADD_get_variant_score fails: CADD PHRED is also available in the dbnsfp block from MyVariant
• AlphaFold prediction unavailable (large proteins >2700aa): Use PDBe_get_uniprot_mappings for experimental structures

---

Special Scenarios

Novel Missense VUS: Check PM5 (other pathogenic at same residue), get AlphaFold2 structure, apply PM1/PP3 as appropriate.

Truncating Variant: Check LOF mechanism, NMD escape, alternative isoforms, ClinGen LOF curation. Apply PVS1 at appropriate strength.

Splice Variant: Run SpliceAI, assess canonical splice distance, in-frame skipping potential. Apply PP3/BP7 based on scores.

---

Output Structure

# Variant Interpretation Report: {GENE} {VARIANT}
## Executive Summary
## 1. Variant Identity
## 2. Population Data
## 3. Clinical Database Evidence
## 4. Computational Predictions
## 5. Structural Analysis
## 6. Literature Evidence
## 7. ACMG Classification
## 8. Clinical Recommendations
## 9. Limitations & Uncertainties
## Data Sources

File naming: {GENE}_{VARIANT}_interpretation_report.md

---

Clinical Recommendations

Pathogenic/Likely Pathogenic: Enhanced screening, risk-reducing options, drug dosing adjustment, reproductive counseling, family cascade screening.

VUS: Do not use for medical decisions. Reinterpret in 1-2 years. Pursue functional studies and segregation data.

Benign/Likely Benign: Not expected to cause disease. No cascade testing needed.

---

Quantified Minimums

Section	Requirement
Population frequency	gnomAD overall + at least 3 ancestry groups
Predictions	At least 3 computational predictors
Literature search	At least 2 search strategies
ACMG codes	All applicable codes listed

---

Cross-Skill References

For amino acid properties at variant position, run: python3 skills/tooluniverse-sequence-analysis/scripts/amino_acids.py --type amino_acid --code X

---

References

• ACMG_CLASSIFICATION.md - Evidence codes, classification algorithm, prediction thresholds, structural/regulatory impact tables
• CODE_PATTERNS.md - Reusable code patterns for each workflow phase
• CHECKLIST.md - Pre-delivery verification
• EXAMPLES.md - Sample interpretations
• TOOLS_REFERENCE.md - Tool parameters and fallbacks