Chromatin Loop Calling (bioSkills)

A Claude Code skill for finding corner-dot chromatin loops in a contact map — and, first, for deciding whether the map is deep enough to look for them at all.

   
Type Claude Skill
Supplier GPTomics bioSkills (community OSS, MIT)
Availability GA — part of the bioSkills collection
Pricing Free / OSS (MIT) — cooltools, chromosight and Mustache are installed separately (all open source)
Capabilities Read/Write — Claude runs the skill’s workflow locally (Python/CLI), not as an MCP tool
Verified works · 2026-08-13
Security cleared · 2026-08-13 — GPTomics/bioSkills MIT, no external credentials

How to install

bioSkills is not an npm package — skills are plain markdown/code read directly by the agent. Clone the repo, then either run the installer for the whole category or copy the single skill directory.

  • Claude Code — clone and install via the bundled script: 
    git clone https://github.com/GPTomics/bioSkills
cd bioSkills
./install-claude.sh --categories "hi-c-analysis"

    The installer copies matching skills into ~/.claude/skills/ (default target). Use ./install-claude.sh --list to preview the skills first.

  • Claude Code / other agents — copy just this one skill: 
    cp -r bioSkills/hi-c-analysis/loop-calling ~/.claude/skills/

    (run from inside your clone — the previous step left you in bioSkills/; otherwise replace bioSkills/ with the absolute path of your clone, e.g. /Users/you/repos/bioSkills).

  • Prerequisites — the core Python stack (required for dots and APA): 
    pip install "cooler>=0.10" "cooltools>=0.7" "bioframe>=0.7"
  • Optional — chromosight 1.6+ (template-correlation calling, and the quantify step used for differential loops): 
    pip install "chromosight>=1.6"

    Confirm with chromosight --version.

  • Optional — Mustache 1.3+ (scale-space multi-resolution detection). Unverified — the skill names Mustache 1.3+ but does not give an install command, and the PyPI distribution name was not confirmed this run; install from the upstream project (ay-lab/mustache) and check mustache --help before use. The cooltools and chromosight paths cover the workflow without it.

What it does

Treats loop calling as a decision about data depth first and an algorithm choice second:

  • Depth assessment — the critical fork. De-novo calling needs roughly 5–10 kb resolution, which in practice means hundreds of millions to billions of valid pairs. Shallower maps are routed to APA on known anchors instead of de-novo calling.
  • De-novo calling — cooltools.dots(), a HiCCUPS-style local-enrichment test using four background donut/neighbourhood models with lambda-chunked FDR.
  • Alternative callers — chromosight (template correlation; also finds borders and stripes) and Mustache (scale-space blob detection).
  • Aggregate peak analysis — cooltools.pileup() builds an APA pileup to confirm a loop set, or to test known CTCF/cohesin anchors in a map too shallow to call de novo.
  • Validation — consensus across callers plus convergent-CTCF support, rather than trusting one caller’s list.
  • Differential loops — union anchor set across conditions, then chromosight quantify to score the same coordinates in every map.

Primary use cases: calling loops or dots from a cooler, deciding de-novo vs APA-on-known-anchors, comparing loops between conditions.

Notes

The scoping rule to read before installing: for protein-anchored assays — HiChIP, PLAC-seq, PCHi-C — this skill explicitly routes you elsewhere (FitHiChIP or MAPS), because dots assumes an all-by-all Hi-C background model that those protocols violate. bioSkills ships a separate hichip-plac-loops skill for that case.

As with the other Hi-C skills, the input cooler must be balanced, and .mcool files must be addressed with a resolution-specific URI (file.mcool::/resolutions/5000) rather than the bare path. Expected-value computation and chromosome-arm regions come from cooltools and bioframe respectively, so arm definitions need chromosome names consistent with the cooler.

Differential loop analysis is deliberately structured as quantify the union anchors in both maps, not intersect two independently-called lists — independent calling near a detection threshold manufactures condition-specific loops that are really just borderline calls.

Upstream skill front-matter name is bio-hi-c-analysis-loop-calling (tool_type: mixed, primary_tool: cooltools); upstream directory hi-c-analysis/loop-calling. Pairs with TAD Detection (domain boundaries rather than focal dots), A/B Compartment Analysis, MACS3 and HOMER for the CTCF/cohesin ChIP-seq anchors used in APA, and bedtools for anchor arithmetic.

Sources

Installed this tool?

Share feedback — install path, OS, errors, workarounds. The form opens with this tool pre-selected and a link back to this page.