Borzoi

SkillMedia

Lets your agent predict how DNA variants and regulatory sequences affect gene expression using genomics models.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Borzoi skill

About this capability

Use Borzoi-style regulatory genomics models for sequence-to-expression or variant-effect analysis. Use when the task asks for noncoding variant impact, regulatory sequence design, or expression prediction.

What this skill tells your AI

The instructions your AI receives, as published by companion-inc/feynman in skills/borzoi/SKILL.md and read by ahel’s review.

Use this skill for regulatory genomics modeling around sequence, variant, and expression effects.

Workflow:

  1. Pin genome build, interval coordinates, reference/alternate alleles, cell type or tissue, strand, and window size.
  2. Confirm the model route: local checkpoint, managed endpoint, notebook package, or external documented service. Record missing checkpoints as setup work, not as a completed model run.
  3. Build input FASTA/variant manifests with source URLs or accession IDs.
  4. Save prediction arrays, summary tables, plots, model version, and runtime metadata as Feynman artifacts.
  5. Compare predicted effects against GTEx, ENCODE, literature, or other source-backed evidence when available.

Keep source-owned genomic coordinates separate from model-owned effect predictions in every output.

Signals

GitHub stars
9k
Forks
1k
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
borzoi
Source
github.com/companion-inc/feynman