Skip to content

Condition Sets (Step 15 of gks_scv_condition_proc)

Overview

Step 15 of the clinvar_ingest.gks_scv_condition_proc procedure assembles individual conditions into structured domain entities for each SCV. SCVs with a single condition produce a Condition record; SCVs with multiple conditions produce a ConditionSet containing a concepts array, conceptSetType, and membershipOperator. The resulting gks_scv_condition_sets table feeds directly into the SCV statement assembly procedure (gks_scv_statement_proc), where it becomes the condition component of the full SCV statement.

Workflow

This step executes as a single query with two CTEs.

Build Individual Condition Records

The enriched_conditions CTE joins each SCV trait from gks_scv_condition_mapping with its normalized trait record from temp_gks_trait, and uses a COUNT(*) OVER window function to classify single vs multi-condition SCVs in one pass. For each condition, the output includes:

  • id — the clinical assertion trait ID (cat_id)
  • name — the CA trait name, falling back to the submitted trait name if the CA name is null
  • conceptType — the CA trait type
  • primaryCoding — from the normalized gks_trait record (MedGen coding)
  • mappings — from the normalized gks_trait record (non-MedGen cross-references)
  • extensions — a concatenation of trait extensions plus condition-specific extensions (see Condition Extensions for full field documentation)

Build Condition Sets for Multi-Condition SCVs

The multi_sets CTE filters to only multi-condition SCVs (where trait_count > 1) and aggregates. The grouping produces:

  • concepts — an array of #/condition/ JSON pointer references to condition records in the condition bundle section
  • conceptSetType — the RCV trait set type (e.g., Disease, Finding)
  • membershipOperator — determines how multiple conditions relate to each other:
  • AND — when the trait relationship type is Finding member or co-occurring condition
  • OR — for all other relationship types

Assemble Final Output

The final query joins temp_gks_scv_trait_sets with the condition and condition set CTEs to produce one row per SCV with two mutually exclusive fields:

Single-condition SCVs populate the condition field:

  • A Condition struct with id, name, conceptType, primaryCoding, mappings, and extensions
  • Extensions include both the individual trait extensions and the trait set extensions
  • If the SCV's submitted trait set type (cats_type) differs from the RCV trait set type, a submittedScvTraitSetType extension is appended

Multi-condition SCVs populate the conditionSet field:

  • A ConditionSet struct with the concepts array, conceptSetType, membershipOperator, and extensions
  • Extensions include the trait set extensions and the optional submittedScvTraitSetType

Output: gks_scv_condition_sets — one row per SCV with either a condition or conditionSet field populated. Pipeline table

Extensions

See Condition Extensions for the complete extension reference, including all individual condition extensions and wrapper-level extensions with examples and custom structure documentation.

Output Tables

Table Description Role
gks_scv_condition_sets Per-SCV condition or condition set with structured codings, mappings, and extensions Pipeline table

Dependencies

  • Source Tables: gks_scv_condition_mapping (persistent), temp_gks_trait (internal), temp_gks_scv_trait_sets (internal)
  • Upstream Steps: Step 1 (Traits), Steps 2–14 (Condition Mapping)
  • Downstream Consumers: gks_scv_statement_proc