Genomics Vocabulary v1.0 Draft (v1.0-draft.0.5)

Overview

Namespace URI:

https://ns.cascadeprotocol.org/genomics/v1#

Preferred Prefix:

genomics:

Version:

1.0-draft.0.5 (2026-08-20)

Status:

Draft

Imports:

cascade: (Core v1), clinical: (Clinical v1)

Schema files:

genomics.ttl (OWL ontology) · genomics.shapes.ttl (SHACL shapes)

The Genomics Vocabulary is a Layer 2 vocabulary that bridges Layer 1 genomics standards (HGVS, ClinVar, VRS, HGNC, Sequence Ontology, HPO, MONDO/OMIM/ORDO, ACMG/AMP via LOINC answer codes, ClinGen review-status taxonomy) to Layer 3 patient-facing genetic counseling summaries authored in checkup:.

Anchored on FHIR Genomics IG profiles (fhir:Observation for variants, haplotypes, and diplotypes; fhir:ServiceRequest for genetic test orders) and on PROV-O for sequencing-run provenance.

What this vocabulary covers

• Variant identity in four parallel languages. HGVS (c./p./g.), ClinVar Variation/RCV IDs, VRS (CAid + VRS object), HGNC gene IDs, Sequence Ontology consequence terms, zygosity, and phase. No single ID is sufficient; storing all of them prevents transcript-version drift on handoff.
• Variant ≠ Interpretation. Two records: a Variant is a molecular observation; a VariantInterpretation is a clinical assertion at a point in time. Reclassifications create a new interpretation linked to the prior via prov:wasRevisionOf (never edit in place).
• Pedigrees. Layer 2 Pedigree + PedigreeMember with HPO phenotype terms, MONDO conditions, per-variant carrier status, and HL7 v3 RoleCode relationship roles. The Layer 3 FamilyHistoryEntry is unchanged.
• Multi-variant units. Haplotype (PharmVar star alleles, HLA typing) and Diplotype (e.g., CYP2D6 *1/*4) anchored on FHIR Genomics IG haplotype/genotype profiles. CopyNumberVariant as a subclass of Variant.
• ClinVar-aligned submitter assertions. SubmitterAssertion + the 7-value ReviewStatus taxonomy (no-assertion through practice-guideline) with star-rating annotations. aggregatedFrom links a VariantInterpretation back to its contributing submitter records (VCV/SCV pattern).
• Genetic test orders + reports. GeneticTestOrder (subclass of fhir:ServiceRequest) with order-status enum; GeneticTest with test-type enum (single-gene, panel, exome, genome, microarray, karyotype, MLPA, repeat-expansion, methylation, RNAseq).
• Sequencing-run provenance. SequencingRun (PROV-O Activity), RawFile as a pointer-and-hash entity (Cascade does not ingest BAM/CRAM bytes), with htsget endpoint, file format, SHA-256 hash, and reference-genome.
• Data-quality tier model. DataQualityTier + four tier individuals (ClinicalGrade, ResearchGrade, ConsumerGrade, UnknownQuality). The SHACL safety constraint requires that any Pathogenic / Likely Pathogenic interpretation either reference a ClinicalGrade Variant or carry requiresConfirmation true.
• Causality vs. classification. interpretationStatus + CausalityStatus enum (Causative / Contributory / UncertainCausality / Rejected) carries Phenopacket-aligned variant-explains-phenotype semantics, distinct from ACMG/AMP classification.

v1-draft.0.2 additions (2026-05-05)

Small additive evolution pass driven by Phase 1 importer gaps. Nothing removed or renamed; existing data continues to validate.

• Generic GeneticTest linkage. genomics:reportedRecord (ObjectProperty, no rdfs:range) covers GeneticTest → non-Variant record references (Diplotypes, Haplotypes, PGx implications). Resolves the HLA tie-break: variantsObserved has rdfs:range genomics:Variant and could not represent these without a range violation.
• VCF-style coordinates. genomics:refAllele, genomics:altAllele, genomics:genomicStartEnd (xsd:string) map LOINC 69547-8 / 69551-0 / 81254-5. Required by the Phase 3 VCF importer and by FHIR Genomics IG variants that lack HGVS.
• Variant origin. genomics:somaticStatus + the SomaticStatus enum (Germline, Somatic, UnknownSomaticStatus) maps LOINC 48002-0. Critical for cancer interpretation and inheritance reasoning.
• Variant allele frequency. genomics:variantAlleleFrequency (xsd:decimal, SHACL-bounded 0.0–1.0) maps LOINC 81258-6. Distinct from mosaicismFraction: VAF is sequencing-evidence; mosaicism is the clinical conclusion.

Key classes

Closed enumerations (named individuals)

• ACMG class: Pathogenic, LikelyPathogenic, VUS, LikelyBenign, Benign (LOINC-coded).
• Inheritance: AutosomalDominant, AutosomalRecessive, XLinkedDominant, XLinkedRecessive, Mitochondrial, YLinked, MultifactorialPolygenic, UnknownInheritance.
• Allelic requirement: Monoallelic, Biallelic.
• Zygosity: Heterozygous, Homozygous, Hemizygous, CompoundHeterozygous, MosaicLow, MosaicHigh.
• Phase: Cis, Trans, PhaseUnknown.
• Carrier status: PositiveAffected, PositiveUnaffected, Negative, NotTested, Obligate.
• Test type: SingleGeneTest, GenePanelTest, ExomeSequencing, GenomeSequencing, ChromosomalMicroarray, KaryotypeAnalysis, MLPA, RepeatExpansionTest, MethylationStudy, RNASequencing.
• Sequencing technology: ShortReadIllumina, LongReadONT, LongReadPacBio, Mixed, GenotypingArray.
• Lab certification: CLIA, CAP, ISO15189, Uncertified.
• File format: BAM, CRAM, FASTQ, gVCF, VCF, BCF, OtherFileFormat.
• Submitter category: SubmitterLaboratory, SubmitterConsortium, SubmitterExpertPanel, SubmitterResearch, SubmitterClinician.
• Order status: OrderPending, OrderInProgress, OrderResulted, OrderCancelled.
• Causality status: Causative, Contributory, UncertainCausality, Rejected.
• Somatic status (v0.2): Germline, Somatic, UnknownSomaticStatus.
• Review status (ClinVar): NoAssertionProvided, CriteriaNotProvided, SingleSubmitter, ConflictingSubmissions, MultipleSubmittersNoConflict, ExpertPanelReviewed, PracticeGuideline.
• Data provenance (genomics-specific): ConsumerArray, ClinicalSequencing, ResearchSequencing, Imported, AIExtractedGenomics.
• Data quality tier: ClinicalGrade, ResearchGrade, ConsumerGrade, UnknownQuality.

Validation (SHACL)

genomics.shapes.ttl defines 12 sh:NodeShape declarations covering every concrete class. Two safety-critical constraints:

• D-Q5 multi-condition cardinality. VariantInterpretation requires exactly one condition and one variantInterpreted. Multi-condition cases produce multiple interpretation instances, each chained to the prior via prov:wasRevisionOf.
• D-QUALITY-TIER safety constraint. An sh:xone on VariantInterpretation: any Pathogenic / Likely Pathogenic classification MUST either reference a ClinicalGrade Variant OR carry requiresConfirmation true. Prevents propagation of high-impact classifications without traceable clinical-grade provenance.

References

• Per-vocab changelog: spec/ontologies/genomics/CHANGELOG.md
• Workstream design package: cascadeprotocol.org/drafts/genomics-v1/ (ONE-PAGER.md, RATIONALE.md, GAP-ANALYSIS.md, CLI-ARCHITECTURE.md, BRCA2 worked example).
• Implementation plan: 05-04-26 Genomics & Advisory IMPLEMENTATION-PLAN.md.
• Schema reference: cascade-protocol-schemas.md (Genomics section).

Provide Feedback

This vocabulary is in draft status. We welcome feedback on terminology fit, clinical workflow alignment, and any gaps for your use case. Submit Feedback