Skillshub methylation-clock
Compute epigenetic age from DNA methylation arrays using PyAging clocks from GEO accessions or local files.
install
source · Clone the upstream repo
git clone https://github.com/ComeOnOliver/skillshub
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/ComeOnOliver/skillshub "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/ClawBio/ClawBio/methylation-clock" ~/.claude/skills/comeonoliver-skillshub-methylation-clock && rm -rf "$T"
manifest:
skills/ClawBio/ClawBio/methylation-clock/SKILL.mdsource content
Methylation Clock
Domain Decisions
Epigenetic age workflows are difficult to reproduce because preprocessing and clock inputs differ across tools and publications. This skill standardizes a PyAging-first pipeline from ingestion to report generation, with explicit reproducibility outputs.
Core Capabilities
- Accepts exactly one input source: GEO accession (
) or local methylation file (--geo-id
).--input - Applies notebook-aligned preprocessing (female derivation and EPICv2 aggregation by default).
- Converts tabular data to AnnData and runs one or more methylation clocks.
- Exports predictions, missing-feature diagnostics, metadata, figures, and reproducibility artifacts.
Input Contract
- Exactly one input source:
- GEO accession with
(example:--geo-id
)GSE139307 - Local file with
(--input
,.pkl
,.pickle
,.csv
,.tsv
,.csv.gz
).tsv.gz
- GEO accession with
- Required output directory via
--output - Optional clock list via
--clocks
Demo And Usage
Demo fixture provenance and checksum are documented in
skills/methylation-clock/data/PROVENANCE.md.
Install optional methylation-clock dependency (not part of the global base requirements):
pip install pyaging>=0.1
# Demo python skills/methylation-clock/methylation_clock.py \ --input skills/methylation-clock/data/GSE139307_small.csv.gz \ --output /tmp/methylation_clock_demo # GEO input python skills/methylation-clock/methylation_clock.py \ --geo-id GSE139307 \ --output /tmp/methylation_clock_geo # Local methylation file python skills/methylation-clock/methylation_clock.py \ --input my_methylation.pkl \ --clocks Horvath2013,AltumAge,PCGrimAge,GrimAge2,DunedinPACE \ --output /tmp/methylation_clock_local
Output Structure
methylation_clock_report/ ├── report.md ├── figures/ │ ├── clock_distributions.png │ └── clock_correlation.png ├── tables/ │ ├── predictions.csv │ ├── prediction_summary.csv │ ├── missing_features.csv │ └── clock_metadata.json └── reproducibility/ ├── commands.sh ├── environment.yml └── checksums.sha256
Safety Rules
- ClawBio is local-first: user methylation data must remain on-device.
- The skill refuses non-empty output directories to avoid silent overwrite.
- Reports must include this disclaimer: "ClawBio is a research and educational tool. It is not a medical device and does not provide clinical diagnoses. Consult a healthcare professional before making any medical decisions."
Agent Boundary
- Route methylation clock requests to
.skills/methylation-clock/methylation_clock.py - Do not infer clinical diagnosis or treatment from clock estimates.
- Trigger terms include: epigenetic age, methylation clock, Horvath, GrimAge, DunedinPACE, GEO, GSE.
- Valid downstream chaining:
for transcriptomic-aging contrasts andrnaseq-de
for cohort context.equity-scorer