Gnomad Database
Google DeepMindOfficialPublished by Google DeepMind
- Skill
- Tool integration
- Science and research
- Apache-2.0
- Low risk
- Has scripts
AutoGineer summary
Written by AutoGineer from the skill’s own files, not by the author
The gnomad-database skill provides a structured interface to query the Genome Aggregation Database (gnomAD). It enables researchers to retrieve allele frequencies for specific genetic variants, obtain gene constraint metrics such as pLI and LOEUF scores, and search for variants within defined genomic regions or genes. The skill includes helper scripts to manage API interactions and rate limits, ensuring consistent data retrieval for genomic analysis.
What it does
- Retrieves global and ancestry-specific allele frequencies for genetic variants.
- Fetches gene constraint metrics including pLI and LOEUF scores.
- Searches for variants within specific genomic regions or genes.
- Supports querying by variant ID, rsID, or gene symbol.
- Provides filtering allele frequency (FAF) data for variant assessment.
When to use it
- When determining the rarity or allele frequency of specific genetic variants.
- When assessing loss-of-function intolerance using gene constraint metrics.
- When identifying variants located within a particular gene or genomic coordinate range.
- When querying structural variants in the context of population genetics.
Advantages
- Automated handling of API rate limits via provided helper scripts.
- Standardized output formats for variant and gene constraint data.
- Integration with local execution environments using managed dependencies.
- Direct access to aggregated population-scale genomic data.
Usage examples
- “Get the allele frequency for variant 1-55516888-G-GA.”
- The agent executes the variant frequency script and returns a JSON file containing global and ancestry-specific allele frequencies for the specified variant.
- “What are the constraint metrics for the BRCA1 gene?”
- The agent retrieves the pLI and LOEUF scores for BRCA1 and saves the data to a JSON file.
- “Find all missense variants in the TP53 gene.”
- The agent queries the database for TP53 variants filtered by the missense consequence and outputs the results to a file.
Install
Download the skill as a ZIP:
https://autogineer.com/agent-skills/google-deepmind/gnomad-database/downloadUnzip it into the cross-client skills folder of your project:
mkdir -p .agents/skills && unzip gnomad-database.zip -d .agents/skills
Documentation by Google DeepMind
gnomAD Database
Prerequisites
uv: Read theuvskill and follow its Setup instructions to ensureuvis installed and on PATH.User Notification: If .licenses/gnomad_database_LICENSE.txt does not already exist in the workspace root directory then (1) prominently notify the user to check the terms at https://gnomad.broadinstitute.org/policies and https://gnomad.broadinstitute.org/data#api, then (2) create the file recording the notification text and timestamp.
Core Rules
Use the Wrapper: ALWAYS execute the provided helper scripts to query the database rather than accessing the database directly. The scripts automatically enforce the gnomAD API rate limits gracefully.
Notification: If this skill is used, ensure this is mentioned in the output.
Utility Scripts
All scripts are located in the scripts/ subdirectory of this skill's
installation directory. When running them, use the full absolute path to the
script (e.g. /path/to/gnomad_database/scripts/get_variant_frequency.py).
1. Variant Frequency. Retrieves global and ancestry-specific allele
frequencies, homozygote counts, and Grpmax Filtering AF (faf95/faf99) for
exome, genome, and total (exome+genome combined) data. The filtering allele
frequency (FAF) is the maximum credible genetic ancestry group AF (lower bound
of the 95% or 99% CI). Variant ID format must be chrom-pos-ref-alt (e.g.,
1-55516888-G-GA). Alternately, you may provide an rsID.
# By variant ID:
uv run scripts/get_variant_frequency.py --variant_id {variant_id} [--dataset {dataset}] --output variant_frequency.json
# By rsID (e.g., rs1800562):
uv run scripts/get_variant_frequency.py --rsid {rsid} [--dataset {dataset}] --output variant_frequency.json2. Gene Constraint. Retrieves constraint metrics for a gene. The response
will explicitly contain pli, and the LOEUF score is represented by
oe_lof_upper.
uv run scripts/get_gene_constraint.py --gene {gene_symbol} --output {gene_symbol}_constraint.json3. Region/Gene Variant Search. Finds all variants in a region or gene.
# By region:
uv run scripts/search_variants.py --chrom {chrom} --start {start} --end {end} --output region_variants.json
# By gene:
uv run scripts/search_variants.py --gene {gene_symbol} --consequence {pLoF|missense} --output {gene_symbol}_variants.jsonReferences
Further documentation on the data: https://gnomad.broadinstitute.org/data#api More general database documentation: https://gnomad.broadinstitute.org/help
Files
- ATTRIBUTION.mdAdded by AutoGineer912 B
- LICENSELicence file added by AutoGineer11.1 KiB
- SKILL.md3.0 KiB
- references/citation.bib5.1 KiB
- scripts/get_gene_constraint.pyScript2.4 KiB
- scripts/get_variant_frequency.pyScript4.2 KiB
- scripts/search_variants.pyScript5.2 KiB
references/citation.bib
@ARTICLE{Karczewski2020,
ABSTRACT = {Genetic variants that inactivate protein-coding genes are a powerful source of information about the phenotypic consequences of gene disruption: genes that are crucial for the function of an organism will be depleted of such variants in natural populations, whereas non-essential genes will tolerate their accumulation. However, predicted loss-of-function variants are enriched for annotation errors, and tend to be found at extremely low frequencies, so their analysis requires careful variant annotation and very large sample sizes1. Here we describe the aggregation of 125,748 exomes and 15,708 genomes from human sequencing studies into the Genome Aggregation Database (gnomAD). We identify 443,769 high-confidence predicted loss-of-function variants in this cohort after filtering for artefacts caused by sequencing and annotation errors. Using an improved model of human mutation rates, we classify human protein-coding genes along a spectrum that represents tolerance to inactivation, validate this classification using data from model organisms and engineered human cells, and show that it can be used to improve the power of gene discovery for both common and rare diseases.},
AUTHOR = {Karczewski, Konrad J. and Francioli, Laurent C. and Tiao, Grace and Cummings, Beryl B. and Alföldi, Jessica and Wang, Qingbo and Collins, Ryan L. and Laricchia, Kristen M. and Ganna, Andrea and Birnbaum, Daniel P. and Gauthier, Laura D. and Brand, Harrison and Solomonson, Matthew and Watts, Nicholas A. and Rhodes, Daniel and Singer-Berk, Moriel and England, Eleina M. and Seaby, Eleanor G. and Kosmicki, Jack A. and Walters, Raymond K. and Tashman, Katherine and Farjoun, Yossi and Banks, Eric and Poterba, Timothy and Wang, Arcturus and Seed, Cotton and Whiffin, Nicola and Chong, Jessica X. and Samocha, Kaitlin E. and Pierce-Hoffman, Emma and Zappala, Zachary and O'Donnell-Luria, Anne H. and Minikel, Eric Vallabh and Weisburd, Ben and Lek, Monkol and Ware, James S. and Vittal, Christopher and Armean, Irina M. and Bergelson, Louis and Cibulskis, Kristian and Connolly, Kristen M. and Covarrubias, Miguel and Donnelly, Stacey and Ferriera, Steven and Gabriel, Stacey and Gentry, Jeff and Gupta, Namrata and Jeandet, Thibault and Kaplan, Diane and Llanwarne, Christopher and Munshi, Ruchi and Novod, Sam and Petrillo, Nikelle and Roazen, David and Ruano-Rubio, Valentin and Saltzman, Andrea and Schleicher, Molly and Soto, Jose and Tibbetts, Kathleen and Tolonen, Charlotte and Wade, Gordon and Talkowski, Michael E. and Aguilar Salinas, Carlos A. and Ahmad, Tariq and Albert, Christine M. and Ardissino, Diego and Atzmon, Gil and Barnard, John and Beaugerie, Laurent and Benjamin, Emelia J. and Boehnke, Michael and Bonnycastle, Lori L. and Bottinger, Erwin P. and Bowden, Donald W. and Bown, Matthew J. and Chambers, John C. and Chan, Juliana C. and Chasman, Daniel and Cho, Judy and Chung, Mina K. and Cohen, Bruce and Correa, Adolfo and Dabelea, Dana and Daly, Mark J. and Darbar, Dawood and Duggirala, Ravindranath and Dupuis, Josée and Ellinor, Patrick T. and Elosua, Roberto and Erdmann, Jeanette and Esko, Tõnu and Färkkilä, Martti and Florez, Jose and Franke, Andre and Getz, Gad and Glaser, Benjamin and Glatt, Stephen J. and Goldstein, David and Gonzalez, Clicerio and Groop, Leif and Haiman, Christopher and Hanis, Craig and Harms, Matthew and Hiltunen, Mikko and Holi, Matti M. and Hultman, Christina M. and Kallela, Mikko and Kaprio, Jaakko and Kathiresan, Sekar and Kim, Bong-Jo and Kim, Young Jin and Kirov, George and Kooner, Jaspal and Koskinen, Seppo and Krumholz, Harlan M. and Kugathasan, Subra and Kwak, Soo Heon and Laakso, Markku and Lehtimäki, Terho and Loos, Ruth J. F. and Lubitz, Steven A. and Ma, Ronald C. W. and MacArthur, Daniel G. and Marrugat, Jaume and Mattila, Kari M. and McCarroll, Steven and McCarthy, Mark I. and McGovern, Dermot and McPherson, Ruth and Meigs, James B. and Melander, Olle and Metspalu, Andres and Neale, Benjamin M. and Nilsson, Peter M. and O'Donovan, Michael C. and Ongur, Dost and Orozco, Lorena and Owen, Michael J. and Palmer, Colin N. A. and Palotie, Aarno and Park, Kyong Soo and Pato, Carlos and Pulver, Ann E. and Rahman, Nazneen and Remes, Anne M. and Rioux, John D. and Ripatti, Samuli and Roden, Dan M. and Saleheen, Danish and Salomaa, Veikko and Samani, Nilesh J. and Scharf, Jeremiah and Schunkert, Heribert and Shoemaker, Moore B. and Sklar, Pamela and Soininen, Hilkka and Sokol, Harry and Spector, Tim and Sullivan, Patrick F. and Suvisaari, Jaana and Tai, E. Shyong and Teo, Yik Ying and Tiinamaija, Tuomi and Tsuang, Ming and Turner, Dan and Tusie-Luna, Teresa and Vartiainen, Erkki and Vawter, Marquis P. and Watkins, Hugh and Weersma, Rinse K. and Wessman, Maija and Wilson, James G. and Xavier, Ramnik J. and {Genome Aggregation Database Consortium}},
URL = {https://doi.org/10.1038/s41586-020-2308-7},
DATE = {2020-05-01},
DOI = {10.1038/s41586-020-2308-7},
ISSN = {1476-4687},
JOURNALTITLE = {Nature},
NUMBER = {7809},
PAGES = {434--443},
TITLE = {The mutational constraint spectrum quantified from variation in 141,456 humans},
VOLUME = {581},
}
Requirements and compatibility
Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta
“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.
Needs: Needs Python
Security
What our scanner noticed· Low risk
- Scripts that your agent may run
This skill includes scripts that your agent may run. Read them before you use it.
A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.
Attribution and licence
- Published by
- Google DeepMind
- Publisher
- Google DeepMind (Official)
- Source
- google-deepmind/science-skills · skills/gnomad_database · commit 4de6786
- Licence
- Apache License 2.0 (redistribution allowed)
- LICENSE (added by AutoGineer from LICENSE)
- ATTRIBUTION.md (written by AutoGineer)
- Synced
AutoGineer is not affiliated with or endorsed by Google DeepMind or Google DeepMind.
Download ZIP
The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download gnomad-database.zip