is a Data Source.
CADD (Combined Annotation Dependent Depletion) scores the deleteriousness of single nucleotide variants and small insertions and deletions throughout the human genome. It integrates many diverse annotations (gene model consequences from Ensembl VEP, conservation, regulatory and epigenetic data, protein-level and splicing predictions) into one metric by training a model to distinguish simulated de novo variants from variants fixed in human populations since the split from the human-chimpanzee ancestor. CADD provides precomputed scores for all possible SNVs in GRCh37 and GRCh38, scored indel sets, a web service for scoring VCF files, a REST API, and offline scoring scripts. Scores are free for non-commercial use; commercial use requires a license from the University of Washington.
| ID | Name | URL | Category | Format | Description |
|---|---|---|---|---|---|
| cadd.website | CADD Website | cadd.gs.washington.edu | Graphical Interface | http | CADD website with single-variant and ... |
| cadd.snvs.grch38 | CADD v1.7 Whole-Genome SNV Scores (GRCh38) | whole_genome_SNVs.tsv.gz (81.5 GB) | Product | tsv | CADD v1.7 precomputed raw and PHRED-s... |
| cadd.snvs.grch37 | CADD v1.7 Whole-Genome SNV Scores (GRCh37) | whole_genome_SNVs.tsv.gz (79.4 GB) | Product | tsv | CADD v1.7 precomputed raw and PHRED-s... |
| cadd.gnomad-indels.grch38 | CADD v1.7 gnomAD v4.0 InDel Scores (GRCh38) | gnomad.genomes.r4.0.indel.tsv.gz (1.2 GB) | Product | tsv | CADD v1.7 scores for small insertions... |
| cadd.downloads | CADD Downloads | download | Product | http | Download page listing all CADD releas... |
| cadd.score | CADD Scoring Web Service | score | Process | http | Web service for scoring uploaded VCF ... |
| cadd.api | CADD API | api | Programming Interface | http | REST API returning CADD raw and PHRED... |
| cadd.scripts | CADD-scripts | CADD-scripts | Process | ❔ | CADD-scripts, a Snakemake-based pipel... |
| cadd.docs | CADD Information | info | Documentation | http | CADD information pages describing the... |
| ID | Name | URL | Category | Format | Relation | Description |
|---|---|---|---|---|---|---|
| myvariant.api | MyVariant.info API | query | Programming Interface | http | had primary source | MyVariant.info REST API (v1) for vari... |
CADD (Combined Annotation Dependent Depletion) is a framework for scoring the deleteriousness of single nucleotide variants (SNVs) and small insertions and deletions (indels) anywhere in the human genome. It was developed at the University of Washington and HudsonAlpha and is now maintained by the Kircher lab at the Berlin Institute of Health at Charite with the University of Washington.
CADD combines many annotations into a single score. The model is trained to separate variants that survived natural selection (fixed in humans since the human-chimpanzee ancestor) from simulated variants that did not face selection. Annotations include Ensembl VEP gene model consequences, conservation scores, regulatory and epigenetic data, splicing predictions and, since v1.7, protein language model and regulatory CNN scores. Each variant gets a raw score and a PHRED-scaled score, where 10 marks the top 10% of all possible SNVs and 20 the top 1%.
CADD scores are freely available for non-commercial applications. Commercial use requires a license from the University of Washington (UW CoMotion).
Created: October 05, 2026 | Last modified: October 05, 2026