Scanpy (single-cell RNA-seq)
暂无简介
- License 许可
- MIT
- 所属仓库
- rhCat/skillChip
- 最近更新
- 2026-07-31
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- Scanpy (single-cell RNA-seq)
- perks
- [ "inspect", "convert", "qc", "preprocess", "reduce_dimensions", "batch_correct", "cluster", "find_markers", "annotate", "score_genes", "pseudobulk", "subset", "plot", "run_pipeline" ]
- skill
- scanpy
02 资源文件 121
- 其他blueprint.drawio2.7 KB
- 参考文档blueprint.json3.6 KB
- 资产文件blueprint.svg8.9 KB
- 参考文档index.json12.8 KB
- 参考文档ledger.json320 B
- 许可证LICENSE.txt1.0 KB
- 参考文档perks.json3.1 KB
- 参考文档perks/annotate/manifesto.json458 B
- 参考文档perks/annotate/metadata.json1.1 KB
- 脚本perks/annotate/src/_common.py4.0 KB
- 脚本perks/annotate/src/annotate.py3.5 KB
- 脚本perks/annotate/src/annotate.sh1.2 KB
- 参考文档perks/annotate/src/contracts.json394 B
- 参考文档perks/annotate/test/case.json253 B
- 参考文档perks/annotate/test/fixture/celltypes.json33 B
- 资产文件perks/annotate/test/fixture/tiny.csv82 B
- 参考文档perks/batch_correct/manifesto.json411 B
- 参考文档perks/batch_correct/metadata.json1.1 KB
- 脚本perks/batch_correct/src/_common.py4.0 KB
- 脚本perks/batch_correct/src/batch_correct.py2.7 KB
- 脚本perks/batch_correct/src/batch_correct.sh1.1 KB
- 参考文档perks/batch_correct/src/contracts.json403 B
- 参考文档perks/batch_correct/test/case.json243 B
- 资产文件perks/batch_correct/test/fixture/tiny.csv82 B
- 参考文档perks/cluster/manifesto.json416 B
- 参考文档perks/cluster/metadata.json967 B
- 脚本perks/cluster/src/_common.py4.0 KB
- 脚本perks/cluster/src/cluster.py2.5 KB
- 脚本perks/cluster/src/cluster.sh1.2 KB
- 参考文档perks/cluster/src/contracts.json321 B
- 参考文档perks/cluster/test/case.json233 B
- 资产文件perks/cluster/test/fixture/tiny.csv82 B
- 参考文档perks/convert/manifesto.json366 B
- 参考文档perks/convert/metadata.json919 B
- 脚本perks/convert/src/_common.py4.0 KB
- 参考文档perks/convert/src/contracts.json311 B
- 脚本perks/convert/src/convert.py1.5 KB
- 脚本perks/convert/src/convert.sh1.0 KB
- 参考文档perks/convert/test/case.json198 B
- 资产文件perks/convert/test/fixture/tiny.csv82 B
- 参考文档perks/find_markers/manifesto.json477 B
- 参考文档perks/find_markers/metadata.json1.1 KB
- 脚本perks/find_markers/src/_common.py4.0 KB
- 参考文档perks/find_markers/src/contracts.json325 B
- 脚本perks/find_markers/src/find_markers.py3.4 KB
- 脚本perks/find_markers/src/find_markers.sh1.3 KB
- 参考文档perks/find_markers/test/case.json234 B
- 资产文件perks/find_markers/test/fixture/tiny.csv82 B
- 参考文档perks/inspect/manifesto.json306 B
- 参考文档perks/inspect/metadata.json896 B
- 脚本perks/inspect/src/_common.py4.0 KB
- 参考文档perks/inspect/src/contracts.json322 B
- 脚本perks/inspect/src/inspect_data.py2.8 KB
- 脚本perks/inspect/src/inspect_data.sh874 B
- 参考文档perks/inspect/test/case.json202 B
- 资产文件perks/inspect/test/fixture/tiny.csv82 B
- 参考文档perks/plot/manifesto.json445 B
- 参考文档perks/plot/metadata.json1.2 KB
- 脚本perks/plot/src/_common.py4.0 KB
- 参考文档perks/plot/src/contracts.json373 B
- 脚本perks/plot/src/plot.py3.5 KB
- 脚本perks/plot/src/plot.sh1.1 KB
- 参考文档perks/plot/test/case.json236 B
- 资产文件perks/plot/test/fixture/tiny.csv82 B
- 参考文档perks/preprocess/manifesto.json565 B
- 参考文档perks/preprocess/metadata.json1.1 KB
- 脚本perks/preprocess/src/_common.py4.0 KB
- 参考文档perks/preprocess/src/contracts.json327 B
- 脚本perks/preprocess/src/preprocess.py3.7 KB
- 脚本perks/preprocess/src/preprocess.sh1.4 KB
- 参考文档perks/preprocess/test/case.json213 B
- 资产文件perks/preprocess/test/fixture/tiny.csv82 B
- 参考文档perks/pseudobulk/manifesto.json395 B
- 参考文档perks/pseudobulk/metadata.json1.2 KB
- 脚本perks/pseudobulk/src/_common.py4.0 KB
- 参考文档perks/pseudobulk/src/contracts.json412 B
- 脚本perks/pseudobulk/src/pseudobulk.py2.9 KB
- 脚本perks/pseudobulk/src/pseudobulk.sh1.1 KB
- 参考文档perks/pseudobulk/test/case.json245 B
- 资产文件perks/pseudobulk/test/fixture/tiny.csv82 B
- 参考文档perks/qc/manifesto.json569 B
- 参考文档perks/qc/metadata.json1.0 KB
- 脚本perks/qc/src/_common.py4.0 KB
- 参考文档perks/qc/src/contracts.json331 B
- 脚本perks/qc/src/qc_analysis.py4.5 KB
- 脚本perks/qc/src/qc_analysis.sh1.4 KB
- 参考文档perks/qc/test/case.json207 B
- 资产文件perks/qc/test/fixture/tiny.csv82 B
- 参考文档perks/reduce_dimensions/manifesto.json525 B
- 参考文档perks/reduce_dimensions/metadata.json1002 B
- 脚本perks/reduce_dimensions/src/_common.py4.0 KB
- 参考文档perks/reduce_dimensions/src/contracts.json325 B
- 脚本perks/reduce_dimensions/src/reduce_dimensions.py2.6 KB
- 脚本perks/reduce_dimensions/src/reduce_dimensions.sh1.3 KB
- 参考文档perks/reduce_dimensions/test/case.json214 B
- 资产文件perks/reduce_dimensions/test/fixture/tiny.csv82 B
- 参考文档perks/run_pipeline/manifesto.json925 B
- 参考文档perks/run_pipeline/metadata.json1.4 KB
- 脚本perks/run_pipeline/src/_common.py4.0 KB
- 参考文档perks/run_pipeline/src/contracts.json326 B
- 脚本perks/run_pipeline/src/run_pipeline.py7.9 KB
- 脚本perks/run_pipeline/src/run_pipeline.sh2.1 KB
- 参考文档perks/run_pipeline/test/case.json213 B
- 资产文件perks/run_pipeline/test/fixture/tiny.csv82 B
- 参考文档perks/score_genes/manifesto.json407 B
- 参考文档perks/score_genes/metadata.json964 B
- 脚本perks/score_genes/src/_common.py4.0 KB
- 参考文档perks/score_genes/src/contracts.json316 B
- 脚本perks/score_genes/src/score_genes.py3.7 KB
- 脚本perks/score_genes/src/score_genes.sh1.1 KB
- 参考文档perks/score_genes/test/case.json253 B
- 参考文档perks/score_genes/test/fixture/signatures.json48 B
- 资产文件perks/score_genes/test/fixture/tiny.csv82 B
- 参考文档perks/subset/manifesto.json426 B
- 参考文档perks/subset/metadata.json1.1 KB
- 脚本perks/subset/src/_common.py4.0 KB
- 参考文档perks/subset/src/contracts.json311 B
- 脚本perks/subset/src/subset.py2.5 KB
- 脚本perks/subset/src/subset.sh1.2 KB
- 参考文档perks/subset/test/case.json229 B
- 资产文件perks/subset/test/fixture/tiny.csv82 B
03 SKILL.md 正文
scanpy — Scanpy (single-cell RNA-seq)
Governed single-cell RNA-seq workflow built on the K-Dense scanpy script toolkit:
inspect, convert, QC, normalize, reduce dimensions, integrate batches, cluster,
find markers, annotate cell types, score gene sets, pseudobulk, subset, and plot.
Each operation is one independent .h5ad-in / artifact-out perk; the perks chain.
Purpose
Run established scanpy operations deterministically and under audit. Every perk
wraps one vendored CLI core (scripts/*.py, sharing _common.py), translates
governed env vars into CLI flags, writes its artifacts under record_store, and
emits one structured-JSON audit line. The heavy science library (scanpy) is the
actual engine; the porter degrades gracefully (still produces its output file and
exits 0) when scanpy is not installed, so governance/contract checks hold offline.
What to look out for
Each tool emits one line of structured JSON (the audit + debug log) and writes its
artifacts under record_store. LOGS TO CHECK: that line + the named report + the
executor run-ledger. Perks that need scanpy installed produce real .h5ad/CSV
artifacts only when the library is present; otherwise a placeholder output file is
written so the contract still passes.
Perks
| perk | tool | nature |
|---|---|---|
inspect |
inspect_data |
read-only / summarize shape, obs/var, layers, embeddings |
convert |
convert |
load any format -> .h5ad |
qc |
qc_analysis |
QC metrics + filter cells/genes + plots, optional Scrublet |
preprocess |
preprocess |
normalize + log1p + HVG, optional scale/regress |
reduce_dimensions |
reduce_dimensions |
PCA + neighbors + UMAP, optional t-SNE |
batch_correct |
batch_correct |
integration: harmony / bbknn / combat |
cluster |
cluster |
Leiden/louvain at one or many resolutions |
find_markers |
find_markers |
rank_genes_groups + per-group CSVs + plots |
annotate |
annotate |
map clusters -> cell types from JSON/CSV mapping |
score_genes |
score_genes |
score gene signatures and/or cell-cycle phase |
pseudobulk |
pseudobulk |
aggregate by sample x cell type -> matrix for pydeseq2 |
subset |
subset |
subset by obs values or gene list |
plot |
plot |
render umap/tsne/pca/violin/dotplot/heatmap/etc. |
run_pipeline |
run_pipeline |
full workflow in one command (counts -> clustered + markers) |
All perks are read-only / file-producing (destructive: false): they read input
files and write new artifacts under record_store; none mutate a remote/live
service. Output is a new .h5ad (or CSV/figures) — inputs are never overwritten.
How to use it
Pick a perk, copy ledger.json -> task-ledger.json, fill its vars + record_store,
then validate -> compose -> compile -> oversight -> executor.
Localized from K-Dense-AI/scientific-agent-skills
scanpy— MIT (see LICENSE.txt).
04 同仓库更多 Skills
Vercel Deploy
暂无简介
Audio Transcribe
暂无简介
Speech (OpenAI TTS)
暂无简介
Skill Installer
暂无简介
demo-skill
A demo skill used as a test fixture for the validate perk. Use when checking that quick_validate.py accepts a well-formed SKILL.md.
demo-skill
A demo skill used as a test fixture for generating agents/openai.yaml. Use when validating the gen_openai_yaml perk.