Data Domains

A data domain is a logical grouping of related data — a family of schemas, their tables, and the owning team. Knowing your question’s domain tells you which schemas to query and which role you need, since access is granted per schema.

What a domain is

Each dedicated domain page covers the data model (schemas, tables, materialization), access and RBAC, and quality rules. General warehouse selection, role hierarchy, and cross-domain query patterns live on the Architecture Overview and Schema Reference pages.

Domains with a dedicated page

• De-Identified Data (DEID) — the de-identified mirror of the production warehouse, with expert-determination transforms and PHI column mapping. Datavant tokenization is documented alongside.
• Clinical LLM Data — LLM-extracted clinical events, materialized as 30 tables in the CLINICAL_LLM schema, including the source LLM_EXTRACTION_RAW table. Semantic views support Cortex Analyst.
• Oncology Products — the Signatera, Altera, and Latitude product bases in LIMS_PUB (MRD / ctDNA, tissue CGP, methylation MRD). Covered on the Oncology Products page.
• Gene Expression (RNA-seq) — Salmon gene / transcript expression facts and run/QC dims in the RNASEQ schema. Covered on the Gene Expression (RNA-seq) page.

Other domains

These are deployed in PROD; refer to the Schema Reference for their tables:

• VARIANTS — genomic variant calls and genetic ethnicity tables. Query patterns are on the Genomics (VARIANTS) Queries page.
• LIMS Foundation & Cohorts — LIMS_PUB is the foundation layer; LIMS_COHORTS is the partner-mart layer.
• ADS Clinical Abstractions — ADS and ADS_STAGING hold clinical abstraction aggregations.
• Salesforce Provider Data — sourced from DATA_SHARE.SALESFORCE, staged into LIMS_PUB.SF_PROVIDER_AGG.
• Kepler Bioinformatics — sourced from DATA_SHARE.KEPLER_DASHBOARD, staged in KEPLER_STAGING.