agentleFS
aws-samplesaws-samples/amazon-bedrock-agents-healthcare-lifesciences/skills/genomics-variant-interpretation/SKILL.md

genomics-variant-interpretation

Skill274 starsChanged 6 months agoMIT-0View on GitHub

Use when interpreting genomic variants from VCF files, performing clinical variant classification using ClinVar/VEP annotations, analyzing allele frequencies against population data (1000 Genomes), or generating clinical reports for genetic counseling.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All queries automatically enforce:

    line 30
  2. Very specific:

    Use analyze_allele_frequencies with 1000 Genomes data:

    line 65
  3. Distinctive:

    Never state a variant is definitively causal without functional evidence

    line 142

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Genomics Variant Interpretation

When to use this skill

  • Interpret variants in specific genes (BRCA1/2, TP53, CYP2D6)
  • Classify variant pathogenicity using ClinVar + VEP annotations
  • Compare allele frequencies against population databases (1000 Genomes)
  • Analyze a patient cohort for pharmacogenomic or cancer-risk variants
  • Generate clinical-grade variant reports

Workflow: Clinical Variant Interpretation

Step 1: Select the appropriate analysis tool

User questionToolParameters
Variants in specific genesquery_variants_by_genegene_symbols, sample_ids
Chromosomal region analysisquery_variants_by_chromosomechromosome, position_range
Rare variant discoveryanalyze_allele_frequenciesfrequency_threshold
Cohort comparisoncompare_sample_variantssample_ids (min 2)
Complex/custom questionsexecute_dynamic_genomics_queryuser_question

Step 2: Apply quality filtering

All queries automatically enforce:

  • qual > 30 (quality score threshold)
  • PASS filter status
  • Cardinality checks on VEP annotation arrays

These filters ensure only high-confidence variants enter clinical interpretation.

Step 3: Interpret clinical significance

Classification hierarchy (act on highest applicable):

ClinVar SignificanceVEP ImpactAction
PathogenicHIGHImmediate clinical attention
PathogenicMODERATEClinical attention, confirm with functional data
Likely_pathogenicHIGHStrong candidate, recommend confirmatory testing
Likely_pathogenicMODERATEMonitor, include in report
Uncertain_significance (VUS)HIGHFlag for reassessment, research interest
VUSMODERATEMonitor, periodic reclassification
Benign / Likely_benignAnyNo clinical action

Priority scoring (used in query results):

Pathogenic + HIGH impact = 10
Pathogenic + MODERATE = 9
Likely_pathogenic + HIGH = 8
Likely_pathogenic + MODERATE = 7
VUS + HIGH = 6
HIGH impact (no ClinVar) = 5
VUS + MODERATE = 4
All others = 1

Step 4: Assess population frequency context

Use analyze_allele_frequencies with 1000 Genomes data:

Frequency categoryThresholdInterpretation
Very Rare< 0.001 (0.1%)Potential novel pathogenic variant
Rare< 0.01 (1%)Candidate for rare disease
Uncommon< 0.05 (5%)May be population-specific
Common>= 0.05Likely benign polymorphism

Rule: Pathogenic variants for Mendelian diseases are almost always < 1% frequency.

Step 5: Generate clinical report

Structure: Patient ID, Gene, Variant (chr:pos:ref>alt), Consequence, Impact, ClinVar significance, Population frequency (1000G AF + rarity category), Associated disease (CLNDN), Clinical interpretation, Recommended follow-up.

Tool Reference

query_variants_by_gene

Input: gene_symbols ("BRCA1,BRCA2,TP53"), sample_ids (optional), include_frequency (bool)
Output: Variants with VEP annotation, ClinVar significance, priority score
Use for: Targeted gene panels, cancer predisposition, pharmacogenomics

query_variants_by_chromosome

Input: chromosome ("17"), sample_ids (optional), position_range ("32000000-33000000")
Output: All PASS variants in region with annotations
Use for: CNV analysis, specific loci investigation, regional patterns

analyze_allele_frequencies

Input: sample_ids (optional), frequency_threshold (default 0.01)
Output: Variants with rarity classification, quality tiers, 1000G comparison
Use for: Rare disease analysis, novel variant discovery, population genetics

compare_sample_variants

Input: sample_ids ("NA21135,NA21137" -- minimum 2)
Output: Per-sample summary: total variants, pathogenic count, impact distribution, quality metrics
Use for: Family studies, cohort stratification, trio analysis

execute_dynamic_genomics_query

Input: user_question (natural language), sample_ids (optional)
Output: Custom SQL generated and executed against HealthOmics stores
Use for: Complex questions not covered by specialized tools

Data Architecture

Variant data lives in AWS HealthOmics stores queried via Athena:

  • Variant Store (genomicsvariantstore): sample, position, alleles, quality, VEP annotations
  • Annotation Store (genomicsannotationstore): ClinVar attributes (CLNSIG, CLNDN, GENEINFO)

Stores are joined on: contigname + start + referenceallele + alternatealleles[1]

VEP annotation fields: symbol, impact, consequence, biotype, sift_prediction, polyphen_prediction ClinVar fields: CLNSIG, CLNDN, GENEINFO, CLNREVSTAT, RS, ALLELEID

Common Analysis Patterns

PatternGenesFilterAction
Cancer predispositionBRCA1, BRCA2, TP53, PALB2, CHEK2, ATMPathogenic/Likely_pathogenicGenetic counseling referral
PharmacogenomicsCYP2D6, CYP2C19, CYP2C9, DPYD, TPMTFunctional impact allelesMedication dosing adjustment
Rare disease triageAll (frequency filter)Very Rare + HIGH + not BenignCandidate list for clinical review

Conventions

  • Lead reports with the most actionable finding first
  • Include population frequency context for every pathogenic call
  • Distinguish germline (inherited) from somatic (tumor) context
  • For VUS: note classification may change with new evidence
  • Never state a variant is definitively causal without functional evidence
  • Include quality metrics (qual, depth) to assess confidence

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.