Skip to content

AlphaGenome Variant Effects

Look up or compute AlphaGenome predicted effects of DNA variants on regulation and splicing

Stars
55,780
Licence
MIT
Files
8 files

What it does

This skill lets an AI agent look up precomputed AlphaGenome Atlas scores for single-letter DNA changes (SNVs) in the human GRCh38 genome, including a ranking score (AVI), a percentile, and 18 attributions explaining what drives each score. It can also run the AlphaGenome model on demand for variants not in the Atlas, compare reference and alternate predicted tracks, and build links to the Atlas website. It is intended for research use only, not clinical decisions.

Great for

  • Bioinformatics researchers prioritising non-coding or regulatory variants from a VCF or credible set
  • Students learning how sequence-to-function models interpret DNA variants
  • Developers building pipelines that call the AlphaGenome API

Skip it if

  • Clinical or diagnostic decisions, since the skill itself says outputs are research-only and not for medical use
  • Indels, rsIDs or GRCh37 coordinates without first normalising them, because the Atlas only covers hg38 SNVs and wrong builds give plausible wrong answers
  • Work that needs a no-key tool or keyless demo, since this needs a DeepMind API key and network access

Try asking

  • “Rank these candidate variants in my VCF by AVI score and keep those with Phred of at least 20”
  • “Show the top feature attribution and strongest track behind chr9:128225994:G>A”
  • “Score this indel in the HBB region with the AlphaGenome model and compare REF and ALT tracks”
  • “Give me Atlas portal links for these variants with the K562 RNA-seq track”
Step-by-step guide: how to use AlphaGenome Variant EffectsInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 8 original files, the MIT licence and an attribution note.

    alphagenome.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip alphagenome.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip alphagenome.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Verify that REF alleles match GRCh38 FASTA before scoring, since the scripts cannot validate the genome build
  • Rank variants by AVI score rather than applying a single strict genome-wide cut-off, as the Atlas guidance recommends
  • Never paste your API key into a command line or commit it to a file; set it as an environment variable

What's inside

  • references/atlas.md
  • references/interpretation.md
  • references/model-api.md
  • scripts/_common.py
  • scripts/atlas_link.py
  • scripts/atlas_query.py
  • scripts/score_variants.py
  • SKILL.md

Redistributed unmodified from K-Dense-AI/scientific-agent-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.