bio-workflows-proteomics-pipeline

name	bio-workflows-proteomics-pipeline
description	End-to-end proteomics workflow from MaxQuant output to differential protein abundance. Orchestrates data import, normalization, imputation, and statistical testing with limma (default) or MSstats for complex feature-level designs. Use when processing mass spectrometry proteomics.
tool_type	mixed
primary_tool	limma
workflow	true
depends_on	["proteomics/data-import","proteomics/proteomics-qc","proteomics/quantification","proteomics/protein-inference","proteomics/differential-abundance"]

Version Compatibility

Reference examples tested with: MSnbase 2.28+, ggplot2 3.5+, limma 3.58+, DEqMS 1.20+, ashr 2.2+

Before using code patterns, verify installed versions match. If versions differ:

• R: packageVersion('<pkg>') then ?function_name to verify parameters

If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying.

Proteomics Pipeline

"Process my proteomics data from raw MS files to differential abundance" → Orchestrate data import (pyopenms/MaxQuant), QC assessment, protein quantification, normalization, differential abundance testing (limma/DEqMS, or MSstats for feature-level designs), and PTM analysis.

Pipeline Overview

Raw MS Data (mzML) ──> MaxQuant/DIA-NN ──> proteinGroups.txt
                                                 │
                                                 ▼
                    ┌────────────────────────────────────────────┐
                    │           proteomics-pipeline              │
                    ├────────────────────────────────────────────┤
                    │  1. Data Import & Filtering                │
                    │  2. Log2 Transform & Normalization         │
                    │  3. Missing Value Imputation               │
                    │  4. QC: PCA, Correlation                   │
                    │  5. Differential Abundance (limma/MSstats) │
                    │  6. Visualization & Export                 │
                    └────────────────────────────────────────────┘
                                                 │
                                                 ▼
                          Differential Proteins + Volcano Plots

Complete R Workflow

library(limma)
library(ggplot2)
library(pheatmap)

# === 1. DATA IMPORT ===
proteins <- read.delim('proteinGroups.txt', stringsAsFactors = FALSE)
cat('Loaded', nrow(proteins), 'protein groups\n')

# Filter contaminants, reverse, only-by-site
proteins <- proteins[proteins$Potential.contaminant != '+' &
                      proteins$Reverse != '+' &
                      proteins$Only.identified.by.site != '+', ]
cat('After filtering:', nrow(proteins), 'proteins\n')

# Extract LFQ intensities
lfq_cols <- grep('^LFQ\\.intensity\\.', colnames(proteins) value  
intensities  proteins lfq_cols
rownamesintensities  proteinsMajority.protein.IDs
colnamesintensities  gsub  colnamesintensities


intensitiesintensities    
log2_int  log2intensities


sample_medians  applylog2_int  median na.rm  
global_median  mediansample_medians
normalized  sweeplog2_int  sample_medians  global_median



valid_rows  rowSumsnormalized  ncolnormalized  
filtered  normalizedvalid_rows 
cat nrowfiltered 


impute_minprob  x 
    nas  x
     nas x
    xnas  rnormnas mean  meanx na.rm      sdx na.rm  
                    sd    sdx na.rm  
    x

imputed  as.data.frametapplyfiltered  impute_minprob



pca  prcomptimputed scale.  
pca_df  data.framePC1  pcax  PC2  pcax  Sample  rownamespcax



sample_info  read.csv
sample_infocondition  factorsample_infocondition

design  model.matrix   condition data  sample_info
colnamesdesign  levelssample_infocondition

fit  lmFitas.matriximputed design
contrast  makeContrastsTreatment  Control levels  design
fit2  contrasts.fitfit contrast
fit2  eBayesfit2 trend   robust  

results  topTablefit2 coef   number   adjust.method  
resultsprotein  rownamesresults
resultssignificant  resultslogFC    resultsadj.P.Val  


cat
cat resultssignificant 
cat resultssignificant  resultslogFC   
cat resultssignificant  resultslogFC   

write.csvresults  row.names  

MSstats Workflow

library(MSstats)

# From MaxQuant
evidence <- read.table('evidence.txt', sep = '\t', header = TRUE)
proteinGroups <- read.table('proteinGroups.txt', sep = '\t', header = TRUE)
annotation <- read.csv('annotation.csv')

# Convert to MSstats format
msstats_input <- MaxQtoMSstatsFormat(evidence = evidence,
                                      proteinGroups = proteinGroups,
                                      annotation = annotation)

# Process data
processed <- dataProcess(msstats_input, normalization = 'equalizeMedians',
                         summaryMethod = 'TMP', censoredInt = 'NA')

# Comparison
comparison <- matrix(  nrow  
rownamescomparison  
colnamescomparison   

results  groupComparisoncontrast.matrix  comparison data  processed

QC Checkpoints

Stage	Check	Action if Failed
Import	>1000 proteins	Re-run MaxQuant
Filter	<30% removed	Check sample prep
Missing	<40% per sample	Check MS performance
PCA	Replicates cluster	Check for batch effects
Stats	>1% differential	Adjust thresholds

Workflow Variants

TMT/iTRAQ Isobaric Labeling

library(MSnbase)

# Load TMT data
tmt_data <- readMSnSet('tmt_psms.txt')

# Normalize with reference channel
tmt_norm <- normalize(tmt_data, method = 'center.median')

# Summarize to protein level
protein_data <- combineFeatures(tmt_norm, groupBy = fData(tmt_norm)$protein, fun = 'median')

# Then proceed with limma as above

SILAC Workflow

# SILAC ratios from MaxQuant
silac <- read.delim('proteinGroups.txt')
ratio_cols <- grep('Ratio.H.L.normalized', colnames(silac), value = TRUE)

# Log2 transform ratios
silac_log2 <- log2(silac[, ratio_cols])

# One-sample t-test against 0 (no change)
results <- apply(silac_log2, 1, function(x) t.test(x, mu = 0)$p.value)

DIA-NN Workflow

# Load DIA-NN report
diann <- read.delim('report.tsv')

# Pivot to matrix
library(tidyr)
protein_matrix <- diann %>%
    select(Protein.Group, Run, PG.MaxLFQ) %>%
    pivot_wider(names_from = Run, values_from = PG.MaxLFQ)

# Then proceed with normalization and limma

Related Skills

• proteomics/data-import - Load MS data formats
• proteomics/proteomics-qc - Quality control before analysis
• proteomics/quantification - Normalization methods
• proteomics/differential-abundance - Statistical testing details
• proteomics/ptm-analysis - Phosphoproteomics and other PTMs
• data-visualization/volcano-and-ma-plots - Volcano plots with LFC shrinkage