# rwdocs — RWD Snowflake Data Warehouse

Internal documentation for the Real-World Data (RWD) Snowflake data warehouse at Natera: access, architecture, querying, and the data catalog. Each linked `.md` file is the full Markdown source of that page, suitable for loading directly into an LLM context.

## Quickstart

- [Introduction](https://rwdocs.natera.com/docs/introduction.md): What the RWD Snowflake warehouse is, its environments and databases, and how to find your path.
- [For Researchers & Analysts](https://rwdocs.natera.com/docs/researchers-analysts.md): Get read-only access and run your first query through Snowsight — Okta SSO, roles, warehouses.
- [For Data Engineers](https://rwdocs.natera.com/docs/data-engineers.md): Programmatic access via SnowSQL, PAT or key-pair auth, dbt setup, and the local development workflow.

## Design & Architecture

- [Architecture Overview](https://rwdocs.natera.com/docs/architecture-overview.md): The layered medallion pipeline: sources → staging → foundation → marts, and the key technical decisions behind it.
- [Dynamic Tables & Refresh](https://rwdocs.natera.com/docs/dynamic-tables.md): How the foundation and mart layer auto-refreshes on a target lag — and the sharp edges worth knowing.
- [PHI / DEID Separation](https://rwdocs.natera.com/docs/phi-deid-separation.md): Identified and de-identified data live in physically separate databases with disjoint role families.
- [Datavant Tokenization](https://rwdocs.natera.com/docs/tokenization.md): In-Snowflake PHI tokenization on Snowpark Container Services, used to join Natera records to partner datasets.

## Querying & Analysis

- [Best Practices & Performance](https://rwdocs.natera.com/docs/best-practices.md): Write efficient queries and pick the right warehouse — clustering, Search Optimization, anti-patterns.
- [Query Cookbook](https://rwdocs.natera.com/docs/query-cookbook.md): Copy-paste query patterns for each major dataset in the CLINICOGENOMICS warehouse.
- [Genomics (VARIANTS) Queries](https://rwdocs.natera.com/docs/genomics-queries.md): Working query patterns for the multi-billion-row variant tables, built around Search Optimization columns.
- [Dashboard & Reporting Queries](https://rwdocs.natera.com/docs/dashboard-queries.md): Querying the DASHBOARD marts and ADS clinical abstraction data.
- [Cortex Intelligence Agent](https://rwdocs.natera.com/docs/cortex-agent.md): Query the warehouse in natural language via the Snowflake Cortex Analyst-powered agent.

## Working in Snowflake

- [Workspaces](https://rwdocs.natera.com/docs/workspaces.md): Snowsight worksheets for interactive SQL and Python (Snowpark) — creating, running, setting role/warehouse context, and sharing.
- [Streamlit in Snowflake](https://rwdocs.natera.com/docs/streamlit-apps.md): Build Python data apps that run inside Snowflake via Snowpark, reading RWD data with the app’s own role and warehouse.
- [Cortex Code (coco)](https://rwdocs.natera.com/docs/cortex-code.md): Snowflake’s agentic AI coding assistant for writing and explaining SQL — available as a desktop CLI and on-platform in Snowsight.

## Data Catalog

- [Data Products & Marketplace](https://rwdocs.natera.com/docs/data-products.md): Two ways to reach RWD data — log into the RWD accounts, or mount a published listing from the Snowflake internal marketplace in your own account.
- [Schema Reference](https://rwdocs.natera.com/docs/schema-reference.md): The map of which schema holds which domain — foundation, cohorts, variants, dashboards, and key join columns.
- [Oncology Products](https://rwdocs.natera.com/docs/oncology.md): The oncology product base tables — Signatera (MRD/ctDNA), Altera (tissue CGP), and Latitude (methylation) in LIMS_PUB, their grain, identity, and key columns.
- [Gene Expression (RNA-seq)](https://rwdocs.natera.com/docs/gene-expression.md): Salmon-quantified RNA-seq gene and transcript expression in the RNASEQ schema — billion-row facts, join keys, TPM vs counts, and query patterns.
- [Data Domains](https://rwdocs.natera.com/docs/data-domains.md): Logical groupings of related data — which domain your question falls into tells you which schemas and role to use.
- [De-Identified Data (DEID)](https://rwdocs.natera.com/docs/deid-data.md): The CLINICOGENOMICS_DEID database — HIPAA Expert Determination transforms, schema mapping, and how to build a DEID model.
- [Annotations](https://rwdocs.natera.com/docs/annotations.md): Reference annotation databases for variant interpretation: gnomAD, ClinVar, CADD, REVEL, AlphaMissense, VEP.

## Engineering & Admin

For builders and administrators of the warehouse.

- [dbt & Build Pipeline](https://rwdocs.natera.com/docs/dbt-workflow.md): How the warehouse is built as dbt models — repository structure, materializations, maturity tags, and what dbt does not manage.
- [Data Tools & Services](https://rwdocs.natera.com/docs/data-tools.md): Supporting tooling: dbt, the Glue Catalog, and the (decommissioned) Vocabulary Service.
- [Snowflake Role Reference](https://rwdocs.natera.com/docs/role-reference.md): The named roles, what each can read and write, the permissions matrix, and warehouse access by role.
- [How to Grant Access](https://rwdocs.natera.com/docs/granting-access.md): The infrastructure-as-code onboarding workflow — declare the user in Terraform, assign roles in RBAC YAML, open an MR.
- [Admin Guide](https://rwdocs.natera.com/docs/admin-guide.md): The administrative governance surface — domain admins, access stewards, and the two systems that manage access.