MCP HubMCP Hub
スキル一覧に戻る

bio-workflows-cytometry-pipeline

majiayu000
更新日 2 days ago
24 閲覧
58
9
58
GitHubで表示
テストtestingautomationdata

について

このスキルは、生のFCSファイルから差動解析に至るまでのフローサイトメトリーおよび質量サイトメトリーデータ処理のためのエンドツーエンドRパイプラインを提供します。CATALYSTとdiffcytを使用して、補正、変換、ゲーティング/クラスタリング、統計的検定を含む完全なワークフローを調整します。サイトメトリーデータ解析のための標準化された自動化パイプラインが必要な場合にご利用ください。

クイックインストール

Claude Code

推奨
プラグインコマンド推奨
/plugin add https://github.com/majiayu000/claude-skill-registry
Git クローン代替
git clone https://github.com/majiayu000/claude-skill-registry.git ~/.claude/skills/bio-workflows-cytometry-pipeline

このコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします

ドキュメント

Flow Cytometry Pipeline

Pipeline Overview

FCS Files ──> Compensation ──> Transformation ──> Gated/Clustered Data
                                                          │
                                                          ▼
                  ┌─────────────────────────────────────────────────┐
                  │            cytometry-pipeline                   │
                  ├─────────────────────────────────────────────────┤
                  │  1. Load FCS Files                              │
                  │  2. Compensation & Transformation               │
                  │  3. QC & Filtering                              │
                  │  4. Clustering (FlowSOM) or Gating              │
                  │  5. Dimensionality Reduction (UMAP)             │
                  │  6. Differential Abundance/State Analysis       │
                  │  7. Visualization                               │
                  └─────────────────────────────────────────────────┘
                                                          │
                                                          ▼
                      Differential Cell Populations + Markers

Complete R Workflow (CATALYST)

library(CATALYST)
library(diffcyt)
library(SingleCellExperiment)
library(flowCore)
library(ggplot2)

# === 1. SETUP PANEL AND METADATA ===
# Panel definition
panel <- data.frame(
    fcs_colname = c('FSC-A', 'SSC-A', 'CD45', 'CD3', 'CD4', 'CD8', 'CD19',
                    'CD14', 'CD56', 'HLA-DR', 'Ki67', 'IFNg'),
    antigen = c('FSC', 'SSC', 'CD45', 'CD3', 'CD4', 'CD8', 'CD19',
                'CD14', 'CD56', 'HLA-DR', 'Ki67', 'IFNg'),
    marker_class = c('none', 'none', 'type', 'type', 'type', 'type', 'type',
                     'type', 'type', 'type', 'state', 'state')
)

# Sample metadata
md <- data.frame(
    file_name = list.files('data/', pattern = '\\.fcs$'),
    sample_id = paste0('Sample', 1:8),
    condition = rep(c('Control', 'Treatment'), each = 4),
    patient_id = rep(paste0('Patient', 1:4), 2)
)

cat('Loading', nrow(md), 'FCS files...\n')

# === 2. LOAD AND PREPARE DATA ===
fcs_files <- file.path('data', md$file_name)
fs <- read.flowSet(fcs_files)

# Apply compensation if stored in FCS
fs_comp <- compensate(fs, spillover(fs[[1]]))

# Prepare SingleCellExperiment with CATALYST
sce <- prepData(fs_comp, panel, md,
                transform = TRUE,
                cofactor = 5,  # For CyTOF use 5, flow cytometry use 150
                FACS = TRUE)

cat('Loaded', ncol(sce), 'cells\n')

# === 3. QC ===
# Per-sample cell counts
table(sce$sample_id)

# Expression distributions
plotExprs(sce, color_by = 'condition')
ggsave('qc_expression_distributions.png', width = 12, height = 8)

# MDS plot for sample similarity
plotMDS(sce, color_by = 'condition')
ggsave('qc_mds.png', width = 8, height = 6)

# === 4. CLUSTERING ===
cat('Clustering...\n')
sce <- cluster(sce,
               features = 'type',  # Use lineage markers
               xdim = 10, ydim = 10,
               maxK = 20,
               seed = 42)

# Metaclustering at different resolutions
table(cluster_ids(sce, 'meta20'))

# === 5. DIMENSIONALITY REDUCTION ===
cat('Running UMAP...\n')
sce <- runDR(sce, dr = 'UMAP', features = 'type')

# Plot UMAP
plotDR(sce, dr = 'UMAP', color_by = 'meta20')
ggsave('umap_clusters.png', width = 8, height = 6)

plotDR(sce, dr = 'UMAP', color_by = 'condition')
ggsave('umap_condition.png', width = 8, height = 6)

# === 6. CLUSTER ANNOTATION ===
# Heatmap of marker expression
plotExprHeatmap(sce, features = 'type', k = 'meta20',
                by = 'cluster_id', scale = 'last', bars = TRUE)
ggsave('heatmap_clusters.png', width = 12, height = 8)

# Manual annotation based on markers
cluster_annotations <- c(
    '1' = 'CD4 T cells',
    '2' = 'CD8 T cells',
    '3' = 'B cells',
    '4' = 'Monocytes',
    '5' = 'NK cells'
    # ... continue for all clusters
)
sce$cell_type <- cluster_annotations[cluster_ids(sce, 'meta20')]

# === 7. DIFFERENTIAL ANALYSIS ===
cat('Running differential analysis...\n')

# Create design matrix
design <- createDesignMatrix(ei(sce), cols_design = 'condition')

# Contrast
contrast <- createContrast(c(0, 1))  # Treatment vs Control

# Differential Abundance (DA)
res_DA <- testDA_edgeR(sce, design, contrast, cluster_id = 'meta20')

da_results <- as.data.frame(rowData(res_DA))
da_results <- da_results[order(da_results$p_adj), ]
cat('\nDifferential Abundance Results:\n')
print(da_results[, c('cluster_id', 'logFC', 'p_val', 'p_adj')])

# Differential State (DS) - marker expression
res_DS <- testDS_limma(sce, design, contrast,
                        cluster_id = 'meta20',
                        markers_include = rownames(sce)[rowData(sce)$marker_class == 'state'])

ds_results <- as.data.frame(rowData(res_DS))
cat('\nDifferential State Results:\n')
sig_ds <- ds_results[ds_results$p_adj < 0.05, ]
print(sig_ds[, c('cluster_id', 'marker_id', 'logFC', 'p_adj')])

# === 8. VISUALIZATION ===
# DA heatmap
plotDiffHeatmap(sce, res_DA, all = TRUE, fdr = 0.05)
ggsave('da_heatmap.png', width = 10, height = 8)

# Abundance boxplots
plotAbundances(sce, k = 'meta20', by = 'cluster_id', group_by = 'condition')
ggsave('abundance_boxplots.png', width = 12, height = 8)

# Volcano plot
da_results$significant <- da_results$p_adj < 0.05
ggplot(da_results, aes(x = logFC, y = -log10(p_adj), color = significant)) +
    geom_point(size = 3) +
    geom_hline(yintercept = -log10(0.05), linetype = 'dashed') +
    scale_color_manual(values = c('gray', 'red')) +
    theme_bw() +
    labs(title = 'Differential Abundance')
ggsave('da_volcano.png', width = 8, height = 6)

# === 9. EXPORT ===
write.csv(da_results, 'da_results.csv', row.names = FALSE)
write.csv(ds_results, 'ds_results.csv', row.names = FALSE)
saveRDS(sce, 'cytometry_analysis.rds')

cat('\nAnalysis complete!\n')
cat('Significant DA clusters:', sum(da_results$p_adj < 0.05), '\n')

flowCore + Manual Gating Workflow

library(flowCore)
library(flowWorkspace)
library(ggcyto)

# Load data
fs <- read.flowSet(list.files('data/', pattern = '\\.fcs$', full.names = TRUE))

# Compensation
comp_matrix <- spillover(fs[[1]])[[1]]
fs_comp <- compensate(fs, comp_matrix)

# Transformation
trans <- estimateLogicle(fs_comp[[1]], colnames(comp_matrix))
fs_trans <- transform(fs_comp, trans)

# Create GatingSet
gs <- GatingSet(fs_trans)

# Apply gates
gs_add_gating_method(gs, alias = 'live',
                     pop = '+', parent = 'root',
                     dims = 'FSC-A,SSC-A',
                     gating_method = 'gate_flowclust_2d',
                     gating_args = list(K = 2, target = c(50000, 25000)))

gs_add_gating_method(gs, alias = 'singlets',
                     pop = '+', parent = 'live',
                     dims = 'FSC-A,FSC-H',
                     gating_method = 'singletGate')

# Visualize gates
autoplot(gs[[1]], 'singlets')

# Extract gated data
gated_data <- gs_pop_get_data(gs, 'singlets')

Python Alternative (FlowCytometryTools)

import flowkit as fk
import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler
from sklearn.cluster import KMeans

# Load FCS files
sample = fk.Sample('sample.fcs')

# Get data as DataFrame
data = sample.as_dataframe(source='raw')

# Compensation (if needed)
comp_matrix = sample.metadata['spill']
data_comp = np.dot(data, np.linalg.inv(comp_matrix))

# Arcsinh transformation
cofactor = 150  # For flow cytometry
data_trans = np.arcsinh(data_comp / cofactor)

# Clustering
scaler = StandardScaler()
data_scaled = scaler.fit_transform(data_trans)

kmeans = KMeans(n_clusters=10, random_state=42)
clusters = kmeans.fit_predict(data_scaled)

QC Checkpoints

StageCheckAction if Failed
LoadingAll FCS files readCheck file integrity
CompensationSpillover values reasonableRecalculate
TransformationDistributions normalizedAdjust cofactor
Events>10K cells per sampleCheck acquisition
Clustering10-30 populationsAdjust K/resolution
DA>3 replicates per groupNeed more samples

Workflow Variants

CyTOF Data

# CyTOF-specific settings
sce <- prepData(fs, panel, md,
                transform = TRUE,
                cofactor = 5,  # CyTOF uses cofactor 5
                FACS = FALSE)  # Not flow cytometry

# Bead normalization should be done upstream (Fluidigm software)

Paired Design

# For paired samples (e.g., pre/post treatment)
design <- createDesignMatrix(ei(sce), cols_design = c('condition', 'patient_id'))

# Include patient as blocking factor
formula <- createFormula(ei(sce), cols_fixed = 'condition', cols_random = 'patient_id')
res_DA <- testDA_voom(sce, formula, contrast)

Related Skills

  • flow-cytometry/fcs-handling - FCS file operations
  • flow-cytometry/compensation-transformation - Data preprocessing
  • flow-cytometry/gating-analysis - Manual gating
  • flow-cytometry/clustering-phenotyping - Unsupervised clustering
  • flow-cytometry/differential-analysis - Statistical testing
  • flow-cytometry/doublet-detection - Remove doublet events
  • flow-cytometry/bead-normalization - CyTOF EQ bead normalization
  • flow-cytometry/cytometry-qc - Comprehensive QC
  • single-cell/clustering - Related clustering methods

GitHub リポジトリ

majiayu000/claude-skill-registry
パス: skills/cytometry-pipeline

関連スキル

content-collections

メタ

This skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.

スキルを見る

evaluating-llms-harness

テスト

This Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.

スキルを見る

sglang

メタ

SGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.

スキルを見る

polymarket

メタ

This skill enables developers to build applications with the Polymarket prediction markets platform, including API integration for trading and market data. It also provides real-time data streaming via WebSocket to monitor live trades and market activity. Use it for implementing trading strategies or creating tools that process live market updates.

スキルを見る