Skip to content
Get packs

Geniml Genomic Interval Toolkit

Validate genomic BED files and plan local Geniml embedding or consensus-universe runs safely

Source Download skill
Repository
K-Dense-AI/scientific-agent-skills
Stars
55,780
Licence
MIT
Files
14 files

What it does

Once installed, the agent follows a safety-first workflow for machine learning on genomic interval sets (BED files of chromosome regions). It checks that coordinates and genome assemblies match, plans Region2Vec or scEmbed training runs, checks model and tokenizer compatibility, and assesses consensus peak universes. The bundled Python scripts only validate and produce plans; they do not run training or contact the internet.

Great for

  • Bioinformatics or computational biology researchers preparing peak files for Geniml embeddings
  • Analysts who need reproducible, checksummed genomic interval pipelines
  • Teams that want a documented compatibility check before running Geniml 0.8.4 with gtars 0.10.0

Skip it if

  • Clinical or patient-identifying decisions: the skill is a research workflow aid and does not validate biological conclusions
  • Users expecting one-click training: the skill plans runs but the user must install packages and execute commands themselves
  • Environments requiring Python versions other than the tested 3.12 stack without independent verification

Try asking

  • “Validate my peaks.bed against the GRCh38 chromosome sizes file before I train anything”
  • “Plan a Region2Vec run on my universe.bed with a bounded number of epochs”
  • “Check whether my saved tokenizer and model artifacts are compatible with geniml 0.8.4”
  • “Assess a consensus peak universe built from my sample manifest using the ccf method”
Step-by-step guide: how to use Geniml Genomic Interval ToolkitInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 14 original files, the MIT licence and an attribution note.

    geniml.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

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

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip geniml.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

  • Always supply the --chrom-sizes file so the validator can check contig names and bounds; without it checks are weaker.
  • Keep the BED input, universe, and model files on local disk and approve any BEDbase or Hugging Face download explicitly.
  • Install into a dedicated project environment with a lockfile, since the full ml extra pulls older pinned packages.

What's inside

  • references/bedspace.md
  • references/consensus_peaks.md
  • references/region2vec.md
  • references/scembed.md
  • references/utilities.md
  • scripts/__init__.py
  • scripts/_common.py
  • scripts/bed_validator.py
  • scripts/consensus_plan.py
  • scripts/corpus_auditor.py
  • scripts/embedding_plan.py
  • scripts/model_artifact_inspector.py
  • scripts/tokenizer_compatibility.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.