Skip to content
Back to skills

572 Usage Guide 8de7bfb9

ASecurity

Unsupervised clustering identifies cell populations without predefined gates. Useful for discovery and high-dimensional CyTOF/spectral flow data.

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 11, 2026
datagobashexpress

Security analysis

A100/100

Scanned October 11, 2026

npx -y skills add tools-only/X-Skills --skill 572-usage-guide_8de7bfb9 --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 572 Usage Guide 8de7bfb9?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for 572 Usage Guide 8de7bfb9
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-572-usage-guide-8de7bfb9/badge)](https://www.skillsdirectory.com/skills/tools-only-572-usage-guide-8de7bfb9)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

SKILL.md
# Clustering and Phenotyping - Usage Guide

## Overview
Unsupervised clustering identifies cell populations without predefined gates. Useful for discovery and high-dimensional CyTOF/spectral flow data.

## Prerequisites
```bash
# R/Bioconductor
BiocManager::install(c('FlowSOM', 'CATALYST', 'Rphenograph'))
```

## Quick Start
Tell your AI agent what you want to do:
- "Cluster my CyTOF data with FlowSOM"
- "Identify cell populations using Phenograph"
- "Annotate clusters based on marker expression"

## Example Prompts
### Clustering
> "Run FlowSOM clustering with 20 metaclusters on my CyTOF data"
> "Apply Phenograph clustering to identify cell populations"
> "Use Leiden clustering via CATALYST"

### Visualization
> "Create a UMAP colored by FlowSOM clusters"
> "Show a heatmap of marker expression per cluster"
> "Generate a cluster abundance plot per sample"

### Annotation
> "Annotate clusters based on canonical marker expression"
> "Merge similar clusters that express the same markers"
> "Create a panel file with type and state marker annotations"

## What the Agent Will Do
1. Prepare SingleCellExperiment or flowSet for clustering
2. Select phenotyping markers (exclude state markers)
3. Apply clustering algorithm (FlowSOM, Phenograph, or Leiden)
4. Generate UMAP embedding
5. Visualize and annotate clusters

## Tips
- FlowSOM is fast and scalable; Phenograph finds rare populations
- Use type markers (lineage) for clustering, state markers for downstream analysis
- Start with more clusters than expected, then merge
- Check delta area plot to help choose number of clusters
- CATALYST provides a streamlined workflow for CyTOF

## Methods Comparison

| Method | Speed | Auto K | Best For |
|--------|-------|--------|----------|
| FlowSOM | Fast | No | Large datasets, defined biology |
| Phenograph | Slow | Yes | Discovery, rare populations |
| Leiden | Medium | Yes | Consistent community detection |

## CATALYST Panel Format
```
fcs_colname,antigen,marker_class
Yb176Di,CD45,type
Er168Di,CD3,type
Nd142Di,Ki67,state
```

## References
- FlowSOM: doi:10.1002/cyto.a.22625
- Phenograph: doi:10.1016/j.cell.2015.05.047
- CATALYST: doi:10.1101/218826

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.