Agent skill
bio-flow-cytometry-fcs-handling
Read and manipulate Flow Cytometry Standard (FCS) files. Covers loading data, accessing parameters, and basic data exploration. Use when loading and inspecting flow or mass cytometry data before preprocessing.
Stars
163
Forks
31
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/fcs-handling
SKILL.md
FCS File Handling
Load FCS Files
r
library(flowCore)
# Read single FCS file
fcs <- read.FCS('sample.fcs', transformation = FALSE, truncate_max_range = FALSE)
# File info
print(fcs)
# Parameter names
colnames(fcs) # Short names
pData(parameters(fcs)) # Full metadata including descriptions
Load Multiple Files
r
# Read multiple files into flowSet
files <- list.files('data', pattern = '\\.fcs$', full.names = TRUE)
fs <- read.flowSet(files, transformation = FALSE, truncate_max_range = FALSE)
# Sample names
sampleNames(fs)
# Access individual samples
fcs1 <- fs[[1]]
Access Expression Data
r
# Get expression matrix
expr <- exprs(fcs)
head(expr)
# Dimensions
dim(expr) # cells x channels
# Channel statistics
summary(expr)
# Get specific channels
cd4_expr <- expr[, 'CD4']
Channel Metadata
r
# Parameter information
params <- pData(parameters(fcs))
print(params)
# Parameter columns:
# - name: short name (e.g., "FL1-A")
# - desc: description (e.g., "CD4")
# - range: max value
# - minRange: min value
# Get channel mapping
channel_map <- setNames(params$desc, params$name)
Rename Channels
r
# Rename using descriptions
rename_channels <- function(fcs) {
params <- pData(parameters(fcs))
new_names <- ifelse(is.na(params$desc) | params$desc == '',
params$name, params$desc)
colnames(fcs) <- new_names
return(fcs)
}
fcs_renamed <- rename_channels(fcs)
Subsetting Data
r
# Subset by cells (rows)
fcs_subset <- fcs[1:1000, ]
# Subset by channels (columns)
fcs_markers <- fcs[, c('CD4', 'CD8', 'CD3')]
# Subset by expression values
high_cd4 <- fcs[exprs(fcs)[, 'CD4'] > 1000, ]
Merge flowSets
r
# Combine multiple flowSets
fs_combined <- rbind2(fs1, fs2)
# Or concatenate into single flowFrame
all_data <- fsApply(fs, exprs)
all_data <- do.call(rbind, all_data)
Write FCS Files
r
# Write single file
write.FCS(fcs, 'output.fcs')
# Write flowSet
write.flowSet(fs, outdir = 'output_dir')
Sample Metadata
r
# Add sample annotations
pData(fs) <- data.frame(
name = sampleNames(fs),
condition = c('Control', 'Control', 'Treatment', 'Treatment'),
patient = c('P1', 'P2', 'P1', 'P2')
)
# Access
pData(fs)
Basic Visualization
r
library(ggcyto)
# Density plot
autoplot(fcs, 'FSC-A')
# Bivariate plot
autoplot(fcs, 'CD4', 'CD8')
# Multiple samples
autoplot(fs, 'CD4', 'CD8')
Check Data Quality
r
# Time parameter check
if ('Time' %in% colnames(fcs)) {
time <- exprs(fcs)[, 'Time']
plot(time, type = 'l', main = 'Acquisition Time')
}
# Event count per file
fsApply(fs, nrow)
# Check for saturated events
saturation <- apply(exprs(fcs), 2, function(x) mean(x == max(x)) * 100)
print(saturation)
Convert to Data Frame
r
# For use with tidyverse
library(tidyverse)
df <- as.data.frame(exprs(fcs))
df$sample <- 'sample1'
# From flowSet
df_all <- fsApply(fs, function(f) {
d <- as.data.frame(exprs(f))
d$sample <- identifier(f)
d
}, simplify = FALSE)
df_all <- bind_rows(df_all)
Related Skills
- compensation-transformation - Apply compensation and transforms
- gating-analysis - Define cell populations
- clustering-phenotyping - Unsupervised analysis
Didn't find tool you were looking for?