Skip to content

Output Reference

The ClinVar-GKS pipeline produces a single compressed JSON file per release, containing all variant representations, clinical classification statements, and supporting reference data. This section documents the output from a consumer perspective — how the file is structured, what each section contains, and how to interpret the data.

For details on how the output is built, see the Pipeline documentation.


Release Format

Each release is available as a gzip-compressed JSON file (.json.gz) containing a root-level object with bundle sections, and optionally as a set of typed Parquet files (one per section). Each bundle section is a keyed collection — the key is the object's unique identifier, and the value is the complete object.

See Output Format Overview for a detailed guide to the bundle structure, reference patterns, and how to navigate between sections.


Sections

Section Content Key Format
sequenceReference VRS sequence references SQ.{digest}
location VRS sequence locations ga4gh:SL.{digest}
allele VRS alleles with expressions ga4gh:VA.{digest}
copyNumberCount Copy number count variants clinvar:{id}
copyNumberChange Copy number change variants clinvar:{id}
gene Gene MappableConcepts ncbigene:{id}
variation Cat-VRS categorical variants clinvar:{variation_id}
condition Trait/disease concepts clinvar.trait:{id}
conditionSet Multi-condition groupings clinvar.traitset:{id}
submitter Submitting organizations clinvar.submitter:{id}
proposition Classification propositions {scv_id}-{CODE} / {vcv_id}-{group}-{prop}-{level}
evidenceLine Evidence lines (SCV, VCV, RCV) {id}.contributing / {scv_id}.{ver}
scv Submitted classifications clinvar.submission:{scv_id}.{version}
vcv Variation-level aggregates {vcv_id}-{group}-{prop}-{level}
rcv Condition-level aggregates {rcv_id}-{group}-{prop}-{level}

Format Conventions

  • Null stripping — null-valued fields and empty arrays/objects are omitted from the output
  • GA4GH identifiers — VRS identifiers use the ga4gh: prefix (e.g., ga4gh:VA.abc123)
  • ClinVar identifiers — ClinVar-scoped identifiers use the clinvar: prefix (e.g., clinvar:12345)
  • JSON pointer references — objects reference each other using #/{section}/{key} strings (e.g., #/allele/ga4gh:VA.abc123)

Cross-Section References

Objects reference each other using #/ JSON pointer strings rather than embedding full objects. This keeps the file compact and avoids duplication. For example:

  • A variation references its allele as #/allele/ga4gh:VA.abc123
  • An SCV statement references its proposition as #/proposition/SCV001234567-PATH
  • A VCV evidence line references its contributing SCVs as #/scv/clinvar.submission:SCV001234567.1

See ID References for the complete reference format guide, identifier patterns, and resolution rules.


Specifications

The output conforms to these GA4GH standards:

  • VRS 2.0 — Variation Representation Specification for allele and copy number representations
  • Cat-VRS — Categorical Variation for grouping variants at a higher categorical level
  • VA-Spec — Variant Annotation Specification for clinical variant statements