# Annotations

The `ANNOTATIONS` database holds reference data for genomic variant interpretation, deployed in all three environments. Annotation runs through VEP-rs (a Rust VEP), producing all-transcripts annotations for the variant pipelines.

## Overview

Annotation reference sets are versioned and live in their own database so the variant pipelines can join calls to interpretation data without bloating `CLINICOGENOMICS`. The annotated variant tables in `VARIANTS` (for example `VARIANT_CALLS_ANNOTATED_OH_GRCH37`) are produced by joining raw calls to these reference sets.

> **Two kinds of 'annotation'**
>
> This page covers **genomic variant annotation** (the reference databases below). For the clinical Annotations application, see the Annotation hub under Data Domains.

## Reference sets

| Source | Purpose |
| --- | --- |
| gnomAD | Population allele frequencies. |
| ClinVar | Clinical significance of variants. |
| CADD | Deleteriousness scoring. |
| REVEL | Missense pathogenicity scoring. |
| AlphaMissense | Deep-learning missense pathogenicity prediction. |
| VEP | Variant Effect Predictor consequence annotations. |