Genomic Variant Annotators

Curated genomic variant annotation modules from the DNA-seq project.

Overview

This dataset contains pre-computed annotation data for genetic variants, organized by module:

Module	Description	Files
longevitymap	Longevity-associated variants	annotations.parquet, studies.parquet, weights.parquet

Schema

annotations.parquet

Variant-level facts linking rsIDs to genes and phenotypes.

• rsid: dbSNP reference ID
• module: Source module name
• gene: Associated gene symbol
• phenotype: Associated phenotype/trait
• category: Functional category

studies.parquet

Per-study evidence from scientific publications.

• rsid: dbSNP reference ID
• module: Source module name
• pmid: PubMed ID
• population: Study population
• p_value: Statistical significance
• conclusion: Study conclusion
• study_design: Type of study

weights.parquet

Curator-defined scoring for variant impact.

• rsid: dbSNP reference ID
• genotype: Genotype as list[str] (e.g., ["C", "T"])
• module: Source module name
• weight: Numeric weight
• state: "protective", "risk", or "neutral"
• priority: Priority level
• conclusion: Curator conclusion
• curator: Curator name
• method: Curation method

Usage

import polars as pl

# Load from HuggingFace
weights = pl.read_parquet("hf://datasets/just-dna-seq/annotators/data/longevitymap/weights.parquet")
studies = pl.read_parquet("hf://datasets/just-dna-seq/annotators/data/longevitymap/studies.parquet")
annotations = pl.read_parquet("hf://datasets/just-dna-seq/annotators/data/longevitymap/annotations.parquet")

Statistics

• Modules: 1 (longevitymap)
• Total files: 3
• Total size: 0.09 MB

License

MIT License - See LICENSE for details.

Citation

If you use this data, please cite the original sources:

• LongevityMap: https://longevitymap.org/