bioinformatics-init-analysis

这段内容是审核时保存的快照。外部来源才是完整且最新的版本。

---
id: bioinformatics-init-analysis
name: bioinformatics-init-analysis
version: 1.0.0
description: |-
  # bioinformatics-init-analysis
stages: ["experiment"]
tools: ["read_file", "search_project", "write_file", "run_terminal"]
summary: |-
  # bioinformatics-init-analysis
primaryIntent: research
intents: ["research"]
capabilities: ["research-planning", "data-processing"]
domains: ["bioinformatics"]
keywords: ["bioinformatics-init-analysis", "experiment"]
source: builtin
status: verified
upstream:
  repo: dr-claw
  path: skills/bioinformatics-init-analysis
  revision: 8322dc4ef575affaa374aa7922c0a0971c6db7d7
resourceFlags:
  hasReferences: true
  hasScripts: true
  hasTemplates: false
  hasAssets: false
  referenceCount: 4
  scriptCount: 10
  templateCount: 0
  assetCount: 0
  optionalScripts: true
---

# bioinformatics-init-analysis

## Canonical Summary

# bioinformatics-init-analysis

## Trigger Rules

Use this skill when the user request matches its research workflow scope. Prefer the bundled resources instead of recreating templates or reference material. Keep outputs traceable to project files, citations, scripts, or upstream evidence.

## Resource Use Rules

- Read from `references/` only when the current task needs the extra detail.
- Treat `scripts/` as optional helpers. Run them only when their dependencies are available, keep outputs in the project workspace, and explain a manual fallback if execution is blocked.

## Execution Contract

- Resolve every relative path from this skill directory first.
- Prefer inspection before mutation when invoking bundled scripts.
- If a required runtime, CLI, credential, or API is unavailable, explain the blocker and continue with the best manual fallback instead of silently skipping the step.
- Do not write generated artifacts back into the skill directory; save them inside the active project workspace.

## Upstream Instructions

# bioinformatics-init-analysis

A Claude Code plugin that automates initial data analysis for high-dimensional single-cell biology data. Supports CyTOF (mass cytometry), scRNA-seq, and flow cytometry with automatic data type detection and plain-language clinical report generation.

## Features

- **7-step pipeline**: Load → QC → Normalize → PCA/UMAP → Cluster → Marker Analysis → Report
- **Auto-detection**: Identifies CyTOF, scRNA-seq, or flow cytometry from file format and marker patterns
- **Clinical reports**: HTML reports with plain-language explanations for medical doctors and non-bioinformaticians
- **Data-type-aware**: QC, normalization, and interpretation adapt to data type
- **Modular**: Run the full pipeline or import individual steps

## Installation

Clone into your Claude Code plugins directory:

```bash
git clone https://github.com/<your-username>/bioinformatics-init-analysis.git \
    ~/.claude/plugins/bioinformatics-init-analysis
```

### Dependencies

```bash
pip install scanpy anndata matplotlib seaborn scipy scikit-learn pandas numpy
# Optional: fcsparser (for .fcs flow cytometry files)
```

## Usage

### As a Claude Code Plugin

Once installed, trigger the skill in Claude Code with phrases like:
- "Run initial analysis on my CyTOF data"
- "QC my single-cell data"
- "Analyze and generate a report for my dataset"

### Command Line

```bash
python3 scripts/run_pipeline.py <input_path> \
    [--data-type auto|cytof|scrnaseq|flow] \
    [--subsample 500] \
    [--output-dir ./analysis_output] \
    [--report-style clinical|technical]
```

### Examples

```bash
# CyTOF directory of CSVs (auto-detected)
python3 scripts/run_pipeline.py /path/to/cytof_csvs/

# scRNA-seq h5ad file with technical report
python3 scripts/run_pipeline.py /path/to/data.h5ad --report-style technical

# Flow cytometry with more cells per sample
python3 scripts/run_pipeline.py /path/to/data.fcs --subsample 2000
```

## Output

```
analysis_output/
├── figures/                    # All generated plots (PNG)
├── processed/
│   └── adata_processed.h5ad   # Processed AnnData