Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

🩸 Rare-Disease Blood RNA-seq Outlier Detection · clawbio | DecimalAI

scienceby clawbio

🩸 Rare-Disease Blood RNA-seq Outlier Detection

clawbio/rare-disease-rnaseq

Source clawbio/clawbio ↗License MIT

Blood RNA-seq expression-outlier detection for rare-disease diagnostics. Cases scored against a control reference panel; outliers ranked and filtered by a haploinsufficient disease-gene panel.

SkillScore59Maintained

Automatic

Clinical / Healthcare pack

Last verified 8/16/2026

What it changes

Passed 10 of 17 test cases with the skill, and 3 of 17 without it.

Without skill18%

With skill59%

Checks it passed that the model alone missed

The default absolute Z-score filtering threshold is 3
The demo generates 100 control samples and 2 case samples
The fallback gene panel comprises 50 haploinsufficient genes

and 7 more in the full report

9 fixed4 passed either way9 still failing0 broke

Pass-rate lift: +41 pp
Tokens: -5%
Agent turns: 0%
Response time: -53%

Model gemini-3.6-flashMeasured 8/16/2026v113 of 22 cases passed with skill · 17 comparable

Tokens, turns and time describe resource use, not a dollar price.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/clawbio-rare-disease-rnaseq/SKILL.md https://app.decimal.ai/s/clawbio-rare-disease-rnaseq/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add clawbio-rare-disease-rnaseq, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

1 · One-time — install the CLI (docs ↗)

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/clawbio-rare-disease-rnaseq/ and pins it in .decimal/skills.lock

decimalai skills add clawbio-rare-disease-rnaseq

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/clawbio-rare-disease-rnaseq/SKILL.md https://app.decimal.ai/s/clawbio-rare-disease-rnaseq/SKILL.md

Raw file: /s/clawbio-rare-disease-rnaseq/SKILL.md ↗ · machine-readable summary at /s/clawbio-rare-disease-rnaseq.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.

1 of 8

Prompt

In which relative directory path are the flagged anomaly calls, along with direction and mechanism annotations, placed?

What was checked, without and with the skill
Checked	Without skill	With skill
The flagged anomalies list is written to tables/outlier_calls.csv	Missed	Passed

case-11 · gemini-3.6-flash · 8/16/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md74 linessingle file

🩸 Rare-Disease Blood RNA-seq Outlier Detection

Reproduces the diagnostic principle of the Genomics England NGRL paper (Blood-based RNA-Seq of 5,412 individuals, medRxiv 2026.03.19.26348811). For each case sample, scores per-gene expression against a control reference panel and flags candidates falling in a curated dosage-sensitive disease-gene panel.

When To Use

A WGS-negative or WGS-VUS rare-disease patient with a paired blood RNA-seq sample
A clinical bioinformatician triaging candidate diagnoses before MDT review
A population-biobank team building an ancestry-matched control reference for outlier calling (e.g. Qatar Biobank for Sidra paediatric cases)

Method

Per-gene robust outlier scoring on log2(CPM+1):

Library-size normalise (CPM), log-transform
For each gene: compute median and MAD across the control panel
For each case-gene cell: modified z = 0.6745 (x − median) / MAD
Flag |z| ≥ threshold (default 3) and gene in disease panel
Rank by |z|, separate down-outliers (haploinsufficiency-consistent) from up-outliers

This implements the diagnostic principle of OUTRIDER (per-gene outlier vs control panel) without the autoencoder, so it runs in seconds with no R/Bioconductor stack. For clinical-grade calls swap to the full DROP pipeline (gagneurlab/drop) which adds OUTRIDER's denoising autoencoder, FRASER2 splicing outliers, and confounder correction. The skill's I/O contract is the same so the upgrade is drop-in.

Input Contract

Counts matrix (.csv or .tsv): rows = genes (HGNC symbol), columns = sample IDs
Cases file (.txt): one case sample ID per line
Controls file (.txt): one control sample ID per line (typically n ≥ 50)
Disease panel (optional, .csv with gene and mechanism columns): defaults to a built-in 50-gene haploinsufficient panel

Output Structure

rdoutlier_report/
├── report.md                     # per-case candidate diagnoses + clinical narrative
├── result.json                   # standard ClawBio envelope
├── figures/
│   └── case_outlier_heatmap.png  # z-scores across cases × top genes
├── tables/
│   ├── outlier_calls.csv         # all flagged outliers with z-score, direction, mechanism
│   └── per_gene_stats.csv        # control median + MAD per gene

Track record

Benchmark

10 / 17

test cases passed, vs 3 / 17 without

Adoption

Users: 0
Times used: 0

Safety and source details

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

clawbio / clawbio

skills/rare-disease-rnaseq/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1

More skills from the same source and nearby in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

See it in action

What's inside

🩸 Rare-Disease Blood RNA-seq Outlier Detection

When To Use

Method

Input Contract

Output Structure

Track record

Benchmark

Adoption

SkillSafety

Source

Related skills