메뉴
XCMS3 workflow for LC-MS/MS metabolomics preprocessing. Covers peak detection, retention time alignment, correspondence (grouping), and gap filling. Use when processing raw LC-MS data into a feature table for untargeted metabolomics.
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/zja2004/BGI-CLI --skill bio-metabolomics-xcms-preprocessing명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | bio-metabolomics-xcms-preprocessing |
| description | XCMS3 workflow for LC-MS/MS metabolomics preprocessing. Covers peak detection, retention time alignment, correspondence (grouping), and gap filling. Use when processing raw LC-MS data into a feature table for untargeted metabolomics. |
| tool_type | r |
| primary_tool | xcms |
Reference examples tested with: MSnbase 2.28+, scanpy 1.10+, xcms 4.0+
Before using code patterns, verify installed versions match. If versions differ:
packageVersion('<pkg>') then ?function_name to verify parametersIf code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying.
Requires Bioconductor 3.18+ with xcms 4.0+ and MSnbase 2.28+.
Goal: Import raw LC-MS files into R for downstream peak detection and alignment.
Approach: Read mzML/mzXML files into an OnDiskMSnExp object using MSnbase for memory-efficient access.
"Process my raw LC-MS data into a feature table" → Detect chromatographic peaks, align retention times across samples, group corresponding peaks, and fill missing values to produce a sample-by-feature intensity matrix.
library(xcms
libraryMSnbase
raw_files list.files pattern full.names
raw_data readMSDataraw_files mode
raw_data
tablemsLevelraw_data
Goal: Attach sample metadata (group labels, injection order) to the raw data object.
Approach: Create a data frame of sample information and assign it to the phenoData slot.
# Sample metadata
sample_info <- data.frame(
sample_name = basename(raw_files),
sample_group = c(rep('Control', 5), rep('Treatment', 5), rep('QC', 3)),
injection_order = 1:length(raw_files)
)
# Assign to phenoData
pData(raw_data) <- sample_info
Goal: Identify chromatographic peaks in centroided LC-MS data.
Approach: Use the CentWave algorithm which detects peaks by continuous wavelet transform on regions of interest defined by m/z and RT.
# CentWave algorithm for centroided data
cwp <- CentWaveParam(
peakwidth = c(5, 30), # Peak width range in seconds
ppm = 15, # m/z tolerance
snthresh = 10, # Signal-to-noise threshold
prefilter = c(3, 1000), # Min peaks and intensity
mzdiff = 0.01, # Minimum m/z difference
noise = 1000, # Noise level
integrate = 1 # Integration method
)
# Run peak detection
xdata <- findChromPeaks(raw_data, param = cwp)
# Summary
head(chromPeaks(xdata))
cat('Peaks found:', nrow(chromPeaks(xdata)), '\n')
Goal: Detect peaks in profile (non-centroided) LC-MS data.
Approach: Use the MatchedFilter algorithm designed for continuum data, which convolves with a Gaussian model peak.
# MatchedFilter for profile/continuum data
mfp <- MatchedFilterParam(
binSize = 0.1,
fwhm = 30,
snthresh = 10,
step = 0.1,
mzdiff = 0.8
)
xdata_profile <- findChromPeaks(raw_data, param = mfp)
Goal: Correct retention time drift across samples to enable peak correspondence.
Approach: Apply Obiwarp alignment which uses dynamic time warping on the TIC profiles to compute sample-wise RT corrections.
# Obiwarp alignment (recommended)
obp <- ObiwarpParam(
binSize = 0.5,
response = 1,
distFun = 'cor_opt',
gapInit = 0.3,
gapExtend = 2.4
)
xdata <- adjustRtime(xdata, param = obp)
# Check alignment
plotAdjustedRtime(xdata)
Goal: Group corresponding chromatographic peaks across samples into consensus features.
Approach: Use peak density-based grouping which models the RT distribution of peaks in m/z slices to identify features present across samples.
# Group peaks across samples
pdp <- PeakDensityParam(
sampleGroups = pData(xdata)$sample_group,
bw = 5, # RT bandwidth
minFraction = 0.5, # Min fraction of samples
minSamples = 1, # Min samples per group
binSize = 0.025 # m/z bin size
)
xdata <- groupChromPeaks(xdata, param = pdp)
# Check feature definitions
featureDefinitions(xdata)
cat('Features:', nrow(featureDefinitions(xdata)), '\n')
Goal: Recover signal for features that were missed during initial peak detection in some samples.
Approach: Integrate intensity in the expected m/z-RT region for features with missing values using ChromPeakAreaParam.
# Fill in missing peaks
fpp <- ChromPeakAreaParam()
xdata <- fillChromPeaks(xdata, param = fpp)
# Alternative: FillChromPeaksParam for more control
fpp2 <- FillChromPeaksParam(
expandMz = 0,
expandRt = 0,
ppm = 0
)
Goal: Generate a samples-by-features intensity matrix with m/z and RT annotations for downstream analysis.
Approach: Extract feature values and definitions from the processed XCMSnExp object and combine into an exportable table.
# Get feature values (intensity matrix)
feature_values <- featureValues(xdata, method = 'maxint', value = 'into')
# Feature definitions (m/z, RT)
feature_defs <- featureDefinitions(xdata)
feature_defs <- as.data.frame(feature_defs)
feature_defs$feature_id <- rownames(feature_defs)
# Combine
feature_table <- cbind(feature_defs[, c('feature_id', 'mzmed', 'rtmed')], feature_values)
rownames(feature_table) <- feature_table$feature_id
# Save
write.csv(feature_table, 'feature_table.csv', row.names = FALSE)
Goal: Assess preprocessing quality through TIC plots, peak counts, RT correction, and PCA.
Approach: Visualize total ion chromatograms, per-sample peak counts, RT adjustment, and PCA of the feature matrix.
# TIC for each sample
tic <- chromatogram(raw_data, aggregationFun = 'sum')
plot(tic)
# Peak count per sample
peak_counts <- table(chromPeaks(xdata)[, 'sample'])
barplot(peak_counts, main = 'Peaks per sample')
# Check RT correction
par(mfrow = c(1, 2))
plotAdjustedRtime(xdata, col = pData(xdata)$sample_group)
# PCA of features
library(pcaMethods)
log_values <- log2(feature_values + 1)
log_values[is.na(log_values)] <- 0
pca <- pca(t(log_values), nPcs = 3, method = 'ppca')
plotPcs(pca, col = as.factor(pData(xdata)$sample_group))
Goal: Identify isotope patterns and adduct groups among detected peaks to reduce feature redundancy.
Approach: Use CAMERA to group peaks by RT correlation, assign isotope clusters, and annotate adduct types.
library(CAMERA)
# Create CAMERA object
xsa <- xsAnnotate(as(xdata, 'xcmsSet'))
# Group by RT
xsa <- groupFWHM(xsa, perfwhm = 0.6)
# Find isotopes
xsa <- findIsotopes(xsa, mzabs = 0.01, ppm = 10)
# Find adducts
xsa <- findAdducts(xsa, polarity = 'positive')
# Get annotated peak list
camera_results <- getPeaklist(xsa)
Goal: Format the XCMS feature table for import into MetaboAnalyst web or R package.
Approach: Transpose the matrix, create M/Z-RT feature names, and prepend sample group information.
# Format for MetaboAnalyst web or R package
export_data <- t(feature_values)
colnames(export_data) <- paste0('M', round(feature_defs$mzmed, 4), 'T', round(feature_defs$rtmed, 1))
# Add sample info
export_df <- data.frame(Sample = rownames(export_data), Group = pData(xdata)$sample_group, export_data)
write.csv(export_df, 'metaboanalyst_input.csv', row.names = FALSE)