gnomAD v4 population variant frequencies via GraphQL API. Allele counts and frequencies stratified by ancestry (AFR, AMR, EAS, NFE, SAS, FIN, ASJ, MID), gene-level constraint (pLI, LOEUF, missense z), and coverage. Identify rare or constrained variants. For clinical pathogenicity use clinvar-database; for GWAS use gwas-database.
SkillScore95Maintained
Automatic
◆Clinical / Healthcare pack
Last verified 8/24/2026
What it changes
Passed 21 of 22 test cases with the skill, and 16 of 22 without it.
Without skill73%
With skill95%
Checks it passed that the model alone missed
Queries dataset gnomad_r4 for gnomAD v4 data retrieval
Uses LOEUF threshold below 0.35 for loss-of-function intolerance
Checks whether genome object is non-null before reading nested fields
and 2 more in the full report
5 fixed16 passed either way1 still failing0 broke
Pass-rate lift
+23 pp
Tokens
+329%
Agent turns
0%
Response time
-26%
Model gemini-3.6-flashMeasured 8/24/2026v121 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 5
Prompt
In gnomAD v4 GraphQL schema, how are gene constraint metrics structured under the gene response object? Provide the field path and names.
What was checked, without and with the skill
Checked
Without skill
With skill
Accesses constraint fields directly under flat gnomad_constraint object
Missed
Passed
case-18 · gemini-3.6-flash · 8/24/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md751 linessingle file
gnomAD Database
Overview
The Genome Aggregation Database (gnomAD) is a resource of aggregated exome and genome sequencing data from 730,000+ individuals. It provides population variant frequencies stratified by 9 ancestry groups, gene-level constraint scores (pLI, LOEUF), and read coverage information. Access is free via a GraphQL API at https://gnomad.broadinstitute.org/api — no authentication required, no official SDK.
When to Use
Checking whether a candidate variant is rare enough to be clinically relevant (AF < 0.1% in all populations)
Retrieving allele frequencies stratified by ancestry group (AFR, AMR, EAS, NFE, SAS, FIN, ASJ, MID) for a variant
Identifying all rare loss-of-function variants in a gene for burden testing or candidate prioritization
Getting gene constraint metrics (pLI, LOEUF) to assess tolerance to loss-of-function variants
Checking read depth coverage for a region to evaluate if low variant frequency reflects low sequencing coverage
Filtering a VCF by population frequency — query gnomAD AF to discard common variants before clinical interpretation
For clinical pathogenicity classifications use clinvar-database; gnomAD provides frequency evidence but does not classify pathogenicity
For GWAS associations at the study level use gwas-database; gnomAD is for population frequency lookups
Prerequisites
Python packages: requests, pandas, matplotlib
Data requirements: gene symbols (e.g., BRCA1), variant IDs (1-69511-A-G format, or rsIDs)
Environment: internet connection; no API key required
Rate limits: no official published limits; use time.sleep(0.5) between requests for polite access; avoid bursts over 10 requests/second
pip install requests pandas matplotlib
Quick Start
import requests
import time
GNOMAD_API = "https://gnomad.broadinstitute.org/api"
def gnomad_query(query: str, variables: dict = None) -> dict:
"""Execute a gnomAD GraphQL query and return the data payload."""
payload = {"query": query, "variables": variables or {}}
r = requests.post(GNOMAD_API, json=payload, timeout=30)