---
title: For Researchers & Analysts
slug: researchers-analysts
section: Quickstart
canonical_url: "https://rwdocs.natera.com/docs/researchers-analysts"
summary: Get read-only access and run your first query through Snowsight — Okta SSO, roles, warehouses.
---

# For Researchers & Analysts

Query the warehouse through Snowsight, the Snowflake web UI. Most analysts never need a CLI or dbt — sign in with Okta SSO, pick a read-only role, and run SQL in the browser.

## What you need

- A Natera email address (`your_name@natera.com`)
- The Snowflake Okta tile provisioned in your Okta dashboard (an admin sets this up — see Step 1)

## Step 1: Request access

Ask in #ndp-rwd-connect-core to start provisioning. An RWD Snowflake admin adds your email to the `ndp-account-request` repo via a merge request. When the MR merges, CI/CD provisions your Snowflake user and the Okta SSO tile automatically.

## Step 2: Pick the right role

Snowflake controls what you can see through your role. The three read roles below inherit downward: a researcher sees everything a viewer sees, and an analyst sees everything a researcher sees.

| Role | Who it is for | What you get |
| --- | --- | --- |
| `RWD_VIEWER_PHI_ROLE` | BI consumers (QuickSight, dashboards) | Read-only access to `LIMS_COHORTS` and `DASHBOARD` only |
| `RWD_RESEARCHER_PHI_ROLE` | Researchers and scientists | Read access to all curated data (foundation, cohorts, dashboards, VARIANTS, seeds) plus the interactive warehouse |
| `RWD_ANALYST_PHI_ROLE` | Power users and CDO analysts | Everything researchers get, plus staging schemas, Cortex AI features, the analytics warehouse, and write access to the RWD_ANALYTICS scratch space |

> **De-identified data**
>
> For de-identified data in `CLINICOGENOMICS_DEID`, the same three tiers exist as DEID roles: `RWD_VIEWER_DEID_ROLE`, `RWD_RESEARCHER_DEID_ROLE`, `RWD_ANALYST_DEID_ROLE`.

## Step 3: Connect via Snowsight

The RWD Snowflake accounts are separate from the main Natera account. If you have only ever used the main account, you will not see them in your account switcher until you sign in at least once.

| Environment | Sign-in URL |
| --- | --- |
| Development | `https://WA16250-NATERA_RWD_DEVELOPMENT.snowflakecomputing.com` |
| Pre-production | `https://WA16250-NATERA_RWD_PREPRODUCTION.snowflakecomputing.com` |
| Production | `https://WA16250-NATERA_RWD_PRODUCTION.snowflakecomputing.com` |

Click the sign-in URL for the environment you need (Production for most analysts), sign in with Okta SSO, and Snowsight caches the account for future visits.

## Step 4: Set your warehouse

A warehouse is the compute Snowflake uses to run your queries. Without one set, queries fail with `No active warehouse selected`. Run this at the start of each session:

```sql
USE WAREHOUSE RWD_PRODUCTION_QUERY_INTERACTIVE_WH;
```

Viewers using the BI role should use the BI warehouse instead:

```sql
USE WAREHOUSE RWD_PRODUCTION_BI_WH;
```

## Your first query

Run this against `LIMS_PUB.BASE_RENASIGHT`, the Renasight (Organ Health) product base table:

```sql
USE ROLE RWD_RESEARCHER_PHI_ROLE;
USE WAREHOUSE RWD_PRODUCTION_QUERY_INTERACTIVE_WH;

SELECT COUNT(DISTINCT casefile_id) AS casefiles
FROM CLINICOGENOMICS.LIMS_PUB.BASE_RENASIGHT;
```

> **Use the product base tables**
>
> Query the per-product base tables — `BASE_RENASIGHT` (Organ Health), and `BASE_SIGNATERA` / `BASE_ALTERA` / `BASE_LATITUDE` (Oncology). The older `BASE_OH` / `BASE_ONC` tables are legacy; don’t build new work on them.

If the query returns a casefile count, you are connected and authorized. To explore partner cohorts, list what is available:

```sql
SHOW TABLES IN SCHEMA CLINICOGENOMICS.LIMS_COHORTS;
```

> **Variant tables are huge**
>
> The `VARIANTS` tables hold billions of rows and need careful query construction. See **Genomics (VARIANTS) Queries** before querying them.

## Next steps

- **Best Practices & Performance** — warehouse selection, clustering, Search Optimization, anti-patterns
- **Architecture Overview** — the schemas and what is in them
- **Schema Reference** — per-table listings and key join columns
- **For Data Engineers** — programmatic access via SnowSQL, dbt, and MCP