SCV Records (Step 1)¶
Overview¶
Step 1 of gks_scv_statement_proc builds the foundational SCV record table temp_gks_scv by extracting and transforming data from scv_summary and clinical_assertion. This table provides the core classification, submitter, and metadata fields consumed by all subsequent steps (2--7).
Transformations¶
Classification and Proposition Type Mapping¶
Each SCV's classification type and statement type are mapped to a proposition type and predicate by joining to clinvar_clinsig_types. This lookup determines how the SCV will be modeled as a VA-Spec proposition -- for example, whether it represents a pathogenicity assertion, an oncogenicity assertion, or a somatic clinical impact assertion.
Evidence Line Target Proposition (Somatic)¶
For somatic clinical impact assertions, the step derives the evidence line target proposition type and predicate based on the assertion's clinical impact type:
- Prognostic -- maps to better outcome or poor outcome predicates
- Diagnostic -- maps to inclusion criteria or exclusion criteria predicates
- Therapeutic -- maps to sensitivity or resistance predicates
Drug Therapy Extraction¶
Drug therapy names are extracted from the somatic clinical impact JSON field of clinical_assertion for therapeutic assertions. These are parsed into an array for use by the target proposition in Step 6.
Assertion Method Parsing¶
Assertion method attributes and citations are parsed from the clinical_assertion record using the parseAttributeSet and parseCitations UDFs. These provide the method name, type, and any linked publications that describe the assertion methodology.
Submission Level¶
The step joins to the submission_level lookup table on rank to resolve the submission level code and label for each SCV. The submission level reflects whether the assertion comes from an expert panel, practice guideline, or other submitter category.
Submitter Struct¶
A submitter struct is constructed with the format clinvar.submitter:{id}, providing a stable identifier for the submitting organization.
Output Fields¶
| Field | Description |
|---|---|
id |
SCV accession identifier |
version |
SCV version number |
proposition_type |
Mapped proposition type from clinvar_clinsig_types |
predicate |
Mapped predicate from clinvar_clinsig_types |
evidence_line_target_proposition |
Target proposition type/predicate for somatic clinical impact assertions |
submitted_date |
Date the SCV was submitted |
created_date |
Date the SCV was created in ClinVar |
evaluated_date |
Date the classification was last evaluated |
classification_name |
Classification label (e.g., Pathogenic, Likely benign) |
classification_code |
Coded classification value |
direction |
supports or disputes |
strength |
Evidence strength (somatic assertions only) |
submitter |
Submitter struct with clinvar.submitter:{id} identifier |
submission_level |
Submission level code |
submission_level_label |
Human-readable submission level label |
drug_therapy |
Array of drug therapy names (somatic therapeutic only) |
assertion_method_attributes |
Parsed assertion method attributes from parseAttributeSet |
assertion_method_citations |
Parsed citations from parseCitations |
Output¶
temp_gks_scv -- one row per SCV with core classification and submitter metadata. Internal
Consumed by Steps 2--7 for qualifier assembly, proposition construction, and final statement assembly.
Dependencies¶
- Source Tables:
scv_summary,clinical_assertion - Lookup Tables:
clinvar_clinsig_types,submission_level - UDFs:
clinvar_ingest.parseAttributeSet,clinvar_ingest.parseCitations