Covariation Analysis (bioSkills)

A Claude Code skill that asks the question most RNA structure claims skip: does the alignment show compensatory substitutions above what phylogeny alone predicts — and does it even have the power to tell?

   
Type Claude Skill
Supplier GPTomics bioSkills (community OSS, MIT)
Availability GA — part of the bioSkills collection
Pricing Free / OSS (MIT) — R-scape is installed separately (open source, Rivas Lab)
Capabilities Read/Write — Claude runs the skill’s workflow locally (Bash/Python), not as an MCP tool
Verified works · 2026-08-10
Security cleared · 2026-08-10 — MIT, provenance matches, R-scape GPLv3

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 "rna-structure"

    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/rna-structure/covariation-analysis ~/.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).

  • Prerequisite — R-scape 2.0+ (the skill drives it as a CLI; it is not bundled): 
    conda install -c conda-forge -c bioconda rscape

    bioconda ships rscape 2.0.4.a (checked 2026-08-08), which satisfies the skill’s 2.0+ requirement. Source tarballs are also published by the Rivas Lab. Confirm with R-scape --version before running the skill.

What it does

Treats “this RNA has a conserved structure” as a hypothesis to be tested, not asserted:

  • Structure testing (-s) — scores the base pairs of a proposed #=GC SS_cons structure against a phylogeny-aware null, and separately scores alternative pairs, so a rejected structure and a differently-paired structure are distinguishable outcomes.
  • Power analysis — estimates the per-pair probability that covariation could have been detected given the alignment’s diversity. This is the step that turns a negative result into a meaningful one.
  • De novo consensus (--cacofold) — builds a covariation-supported structure when there is no trusted structure to test, suitable for seeding a covariance model or a restrained fold.
  • Outputs — <msa>.cov (covarying pairs with position, score, E-value, substitutions, power), <msa>.power, <msa>.sorted.cov, and an R2R structure diagram as .svg/.pdf.

Working numbers carried by the skill: default E-value target 0.05; a mean alignment power of roughly 10% is the line between “cannot infer” and “rejects the structure”; alignments should be diverse — around 60% average pairwise identity rather than 90–95% — and it is the number of independent substitutions, not the raw sequence count, that buys power.

Primary use cases: validating a conserved-structure claim before building on it, deciding whether an alignment can test structure at all, generating a covariation-supported consensus.

Notes

Two rules do the real work here. First, a low-power negative says nothing — it is not evidence against the structure, and the skill refuses to report it as such. Second, a raw positive covariation score is not enough; only covariation above the phylogenetic null counts, which is the whole reason R-scape exists rather than a plain mutual-information calculation. The skill also draws a boundary readers often blur: covariation tests whether a structure is conserved, which is a different question from whether the transcript is real, expressed, or functional — that needs expression and functional evidence.

The canonical application is the negative result: R-scape found no covariation support for the proposed HOTAIR, Xist, and SRA lncRNA structures, and the skill uses that as its worked example of an adequately-powered rejection.

Input must be a deep, diverse Stockholm-format alignment carrying a #=GC SS_cons line. Upstream skill front-matter name is bio-rna-structure-covariation-analysis; upstream directory rna-structure/covariation-analysis. Pairs naturally with ncRNA Search (a CaCoFold consensus can seed a covariance model), ViennaRNA (thermodynamic folding, which covariation validates rather than replaces), RNA Structure Probing (experimental rather than evolutionary evidence), and Rfam for curated family alignments to test against.

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.