AI Skill MartBrowse skillsPromptsProgramsLearnWhat is a Skill?
 
← Back to list

Gtars Genomic Interval Analysis

A skill for doing genomic interval set algebra, coverage, tokenization, and refget work with gtars (Python/Rust/CLI) under exact version pins and explicit safety gates.

Data & AnalyticsAdvanced★ 33,030⑂ 3,248AI score 8/10Last updated: Aug 9, 2026

What it does

  • Pins down the actual gtars 0.9.x API surface (Python 0.9.2, Rust/CLI 0.9.0) so the assistant stops guessing function names.
  • Enforces a genomic data contract before any operation: 0-based half-open BED coordinates, u32 bounds, exact contig-name matching, sorting and strand policies.
  • Shows worked examples of count_overlaps, find_overlaps, intersect_all, jaccard, reduce, setdiff, and consensus.
  • Puts an explicit approval gate in front of anything with network or disk side effects (pretrained tokenizers, remote refget, BEDbase cache, RegionSet(url) behavior).
  • Ships six deterministic local inspector CLIs (bed_validator, execution_plan, coverage_preflight, and more) for validation and dry-run planning.
  • Lists removed/stale API forms (gtars.RegionSet, TreeTokenizer, uniwig generate, etc.) so old snippets don't get pasted back in.

Who it's for

  • Bioinformaticians and analysis engineers working with BED / fragment files
  • ATAC-seq, ChIP-seq, and single-cell fragment pipeline maintainers
  • Teams burned before by assembly mixups, chr prefixes, or unstated sort assumptions
  • ML dataset builders who need patient-level splits and leakage control

Example uses

  1. "Validate this peak BED against GRCh38 chrom.sizes and tell me if it's sorted" → runs scripts/bed_validator.py and summarizes offending coordinates.
  2. "Give me per-region overlap counts and the jaccard index between query and universe" → generates Python using count_overlaps / jaccard with correct submodule imports.
  3. "I want a pretrained Hugging Face tokenizer" → the skill first demands network approval, an immutable revision, and checksum verification, otherwise falls back to local Tokenizer.from_bed.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/K-Dense-AI/scientific-agent-skills/HEAD/skills/gtars/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on AI Skill Mart.
Copy the skills/gtars folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/gtars/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/K-Dense-AI/scientific-agent-skills.git && mkdir -p ~/.claude/skills && cp -r scientific-agent-skills/skills/gtars ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repository: git clone https://github.com/K-Dense-AI/scientific-agent-skills.git
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy the skill, keeping its bundled folders: cp -r scientific-agent-skills/skills/gtars ~/.claude/skills/
  5. Verify with ls ~/.claude/skills/gtars — you should see SKILL.md, references/, and scripts/.
  6. Restart Claude Code and try a prompt like "analyze overlaps in this BED file".
  7. To actually run gtars, create an isolated environment: uv venv --python 3.11 .venv-gtars, then uv pip install --python .venv-gtars/bin/python "gtars==0.9.2".
  8. Only enable network-capable features (pretrained tokenizers, remote refget, BEDbase caching) after you explicitly approve the host and data, as the skill instructs.
View source on GitHub ↗License: MIT