Pyopenms Mass Spectrometry

uv pip install pyopenms numpy pandas matplotlib

decisions:
  - id: D1
    param: instrumentResolution
    kind: required
    source: data
    ask: "Is this high-resolution data, and are the spectra already centroided or still in profile mode?"
    default: null

  - id: D2
    param: signalToNoiseThreshold
    kind: required
    source: user
    depends_on: [D1]
    ask: "How far above noise must a peak rise to be kept during centroiding?"
    default: 1.0
    skip_if: "spectra already centroided by the acquisition software"

  - id: D3
    param: massTolerance
    kind: required
    source: data
    depends_on: [D1]
    ask: "What mass accuracy should feature detection and linking assume?"
    default: "10 ppm"

  - id: D4
    param: retentionTimeTolerance
    kind: required
    source: data
    ask: "How far apart in retention time may the same feature appear across runs?"
    default: "100 s"

  - id: D5
    param: chargeStateRange
    kind: required
    source: user
    ask: "Which charge states should features be searched for?"
    default: "1 to 3 - widen for intact protein or metabolite work"

  - id: D6
    param: intensityNormalization
    kind: optional
    source: user
    ask: "Should spectra be normalized before comparison, and to total ion current or to the base peak?"
    default: "not normalized"

  - id: D7
    param: smoothingFilter
    kind: optional_conditional
    source: data
    ask: "Does the signal need smoothing before peak picking, and with which filter width?"
    default: "no smoothing"

import pyopenms as ms

# Load mzML file
exp = ms.MSExperiment()
ms.MzMLFile().load("sample.mzML", exp)
print(f"Spectra: {exp.getNrSpectra()}, Chromatograms: {exp.getNrChromatograms()}")

# Examine first spectrum
spec = exp.getSpectrum(0)
mz, intensity = spec.get_peaks()
print(f"MS level: {spec.getMSLevel()}, RT: {spec.getRT():.2f}s, Peaks: {len(mz)}")

# Quick preprocessing: smooth + centroid
gauss = ms.GaussFilter()
p = gauss.getParameters(); p.setValue("gaussian_width", 0.1); gauss.setParameters(p)
gauss.filterExperiment(exp)

picker = ms.PeakPickerHiRes()
centroided = ms.MSExperiment()
picker.pickExperiment(exp, centroided)
print(f"Centroided spectra: {centroided.getNrSpectra()}")

import pyopenms as ms

# Read mzML (standard MS format)
exp = ms.MSExperiment()
ms.MzMLFile().load("data.mzML", exp)

# Indexed access for large files (memory-efficient)
loader = ms.IndexedMzMLFileLoader()
indexed_file = ms.OnDiscMSExperiment()
loader.load("large_data.mzML", indexed_file)
spec = indexed_file.getSpectrum(0)  # Load single spectrum on demand
print(f"Total spectra: {indexed_file.getNrSpectra()}")

# Read identification results (idXML)
protein_ids, peptide_ids = [], []
ms.IdXMLFile().load("results.idXML", protein_ids, peptide_ids)
print(f"Peptide IDs: {len(peptide_ids)}, Protein IDs: {len(protein_ids)}")

# Read feature map
fm = ms.FeatureMap()
ms.FeatureXMLFile().load("features.featureXML", fm)
print(f"Features: {fm.size()}")

# Write mzML with compression
exp_out = ms.MSExperiment()
# ... populate experiment ...
ms.MzMLFile().store("output.mzML", exp_out)

# Read FASTA database
entries = []
ms.FASTAFile().load("database.fasta", entries)
print(f"Proteins in DB: {len(entries)}")
for e in entries[:3]:
    print(f"  {e.identifier}: {e.sequence[:30]}...")

import pyopenms as ms

exp = ms.MSExperiment()
ms.MzMLFile().load("raw.mzML", exp)

# Gaussian smoothing
gauss = ms.GaussFilter()
p = gauss.getParameters()
p.setValue("gaussian_width", 0.15)  # m/z width
gauss.setParameters(p)
gauss.filterExperiment(exp)

# Savitzky-Golay smoothing (alternative)
sg = ms.SavitzkyGolayFilter()
p = sg.getParameters()
p.setValue("frame_length", 15)  # Must be odd
sg.setParameters(p)
# sg.filterExperiment(exp)  # Use one smoother, not both

# Peak picking (centroiding) — required before feature detection
picker = ms.PeakPickerHiRes()
p = picker.getParameters()
p.setValue("signal_to_noise", 1.0)
picker.setParameters(p)
centroided = ms.MSExperiment()
picker.pickExperiment(exp, centroided)
print(f"Raw peaks in spec 0: {exp.getSpectrum(0).size()}")
print(f"Centroided peaks: {centroided.getSpectrum(0).size()}")

# Normalization
normalizer = ms.Normalizer()
p = normalizer.getParameters()
p.setValue("method", "to_one")  # "to_one" or "to_TIC"
normalizer.setParameters(p)
normalizer.filterPeakMap(centroided)

# Peak filtering — remove low-intensity noise
mower = ms.ThresholdMower()
p = mower.getParameters()
p.setValue("threshold", 100.0)  # Minimum intensity
mower.setParameters(p)
mower.filterPeakMap(centroided)

# Baseline reduction
morph = ms.MorphologicalFilter()
p = morph.getParameters()
p.setValue("struc_elem_length", 3.0)  # m/z window
morph.setParameters(p)
morph.filterExperiment(exp)

import pyopenms as ms

# Load centroided data
exp = ms.MSExperiment()
ms.MzMLFile().load("centroided.mzML", exp)

# Feature detection (proteomics — centroided data)
ff = ms.FeatureFinder()
features = ms.FeatureMap()
seeds = ms.FeatureMap()
params = ms.FeatureFinder().getParameters("centroided")
ff.run("centroided", exp, features, params, seeds)
print(f"Detected {features.size()} features")

# Access feature properties
for f in features[:5]:
    print(f"  RT: {f.getRT():.1f}s, m/z: {f.getMZ():.4f}, "
          f"intensity: {f.getIntensity():.0f}, quality: {f.getOverallQuality():.3f}")

# Feature linking across samples — align retention times first
aligner = ms.MapAlignmentAlgorithmPoseClustering()
p = aligner.getParameters()
p.setValue("max_num_peaks_considered", 1000)
aligner.setParameters(p)

# Link features into consensus map
linker = ms.FeatureGroupingAlgorithmQT()
p = linker.getParameters()
p.setValue("distance_RT:max_difference", 60.0)  # seconds
p.setValue("distance_MZ:max_difference", 10.0)   # ppm
linker.setParameters(p)

consensus = ms.ConsensusMap()
linker.group([features_sample1, features_sample2, features_sample3], consensus)
print(f"Consensus features: {consensus.size()}")

# Export to pandas for downstream analysis
import pandas as pd
df = consensus.get_df()
print(f"Consensus table: {df.shape}")

import pyopenms as ms

# Load search engine results
protein_ids, peptide_ids = [], []
ms.IdXMLFile().load("search_results.idXML", protein_ids, peptide_ids)

# Examine peptide hits
for pep_id in peptide_ids[:3]:
    print(f"Spectrum: RT={pep_id.getRT():.1f}, MZ={pep_id.getMZ():.4f}")
    for hit in pep_id.getHits():
        seq = hit.getSequence()
        print(f"  {seq} score={hit.getScore():.4f} charge={hit.getCharge()}")

# FDR filtering (target-decoy approach)
fdr = ms.FalseDiscoveryRate()
fdr.apply(peptide_ids)

# Filter at 1% FDR
filtered = []
for pep_id in peptide_ids:
    hits = [h for h in pep_id.getHits() if h.getScore() <= 0.01]
    if hits:
        pep_id.setHits(hits)
        filtered.append(pep_id)
print(f"Peptide IDs at 1% FDR: {len(filtered)}")

# Protein inference
inference = ms.BasicProteinInferenceAlgorithm()
inference.run(peptide_ids, protein_ids)

for prot_id in protein_ids:
    for hit in prot_id.getHits()[:5]:
        print(f"Protein: {hit.getAccession()}, score: {hit.getScore():.4f}")

# Peptide sequence handling
seq = ms.AASequence.fromString("PEPTIDER")
print(f"Molecular weight: {seq.getMonoWeight():.4f}")
print(f"Formula: {seq.getFormula()}")

# Modified sequence
mod_seq = ms.AASequence.fromString("PEPTM(Oxidation)DER")
print(f"Modified weight: {mod_seq.getMonoWeight():.4f}")

# Enzymatic digestion
digestor = ms.ProteaseDigestion()
digestor.setEnzyme("Trypsin")
digest = []
digestor.digest(ms.AASequence.fromString("MKWVTFISLLLLFSSAYSRGVFRR"), digest)
print(f"Tryptic peptides: {len(digest)}")

import pyopenms as ms

# Step 1: Load and centroid raw data
exp = ms.MSExperiment()
ms.MzMLFile().load("metabolomics_sample.mzML", exp)

picker = ms.PeakPickerHiRes()
centroided = ms.MSExperiment()
picker.pickExperiment(exp, centroided)

# Step 2: Feature detection for metabolomics (small molecules)
ff = ms.FeatureFinder()
features = ms.FeatureMap()
seeds = ms.FeatureMap()
params = ms.FeatureFinder().getParameters("centroided")
params.setValue("isotopic_pattern:charge_low", 1)
params.setValue("isotopic_pattern:charge_high", 3)
ff.run("centroided", centroided, features, params, seeds)
print(f"Detected features: {features.size()}")

# Step 3: Adduct detection (group related adducts)
decharger = ms.MetaboliteAdductDecharger()
p = decharger.getParameters()
p.setValue("potential_adducts", "H:+:0.6;Na:+:0.3;K:+:0.1")  # Positive mode
decharger.setParameters(p)
# decharger.compute(features, feature_map_out, consensus_map_out)

# Step 4: RT alignment across samples
import pyopenms as ms
import pandas as pd

# Assuming feature maps from multiple samples
sample_files = ["sample1.featureXML", "sample2.featureXML", "sample3.featureXML"]
feature_maps = []
for f in sample_files:
    fm = ms.FeatureMap()
    ms.FeatureXMLFile().load(f, fm)
    feature_maps.append(fm)

# Align retention times
aligner = ms.MapAlignmentAlgorithmPoseClustering()
aligner.setReference(0)  # Use first sample as reference

# Step 5: Link features to consensus map
linker = ms.FeatureGroupingAlgorithmQT()
p = linker.getParameters()
p.setValue("distance_RT:max_difference", 30.0)
p.setValue("distance_MZ:max_difference", 10.0)
linker.setParameters(p)

consensus = ms.ConsensusMap()
linker.group(feature_maps, consensus)

# Step 6: Export metabolite table
df = consensus.get_df()
print(f"Feature table: {df.shape[0]} features × {df.shape[1]} columns")
print(df[["RT", "mz", "intensity_0", "intensity_1", "intensity_2"]].head())

algo = ms.AlgorithmClass()           # 1. Instantiate
params = algo.getParameters()         # 2. Get parameters
params.setValue("param_name", value)   # 3. Set values
algo.setParameters(params)            # 4. Apply
# algo.process(input, output)         # 5. Execute

for key in params.keys():
    print(f"{key}: {params.getValue(key)} (type: {params.getDescription(key)})")

import pyopenms as ms

# Load → Smooth → Centroid → Detect → Export
exp = ms.MSExperiment()
ms.MzMLFile().load("proteomics.mzML", exp)

# Preprocessing
gauss = ms.GaussFilter()
p = gauss.getParameters(); p.setValue("gaussian_width", 0.1); gauss.setParameters(p)
gauss.filterExperiment(exp)

picker = ms.PeakPickerHiRes()
centroided = ms.MSExperiment()
picker.pickExperiment(exp, centroided)

# Feature detection
ff = ms.FeatureFinder()
features = ms.FeatureMap()
params = ff.getParameters("centroided")
ff.run("centroided", centroided, features, params, ms.FeatureMap())

# Export to pandas
df = features.get_df()
print(f"Features: {df.shape}")
df.to_csv("proteomics_features.csv", index=False)

import pyopenms as ms
import pandas as pd

# Load and filter identifications
protein_ids, peptide_ids = [], []
ms.IdXMLFile().load("search.idXML", protein_ids, peptide_ids)

# FDR filtering
fdr = ms.FalseDiscoveryRate()
fdr.apply(peptide_ids)

# Build results table
rows = []
for pep_id in peptide_ids:
    for hit in pep_id.getHits():
        if hit.getScore() <= 0.01:  # 1% FDR
            rows.append({
                "sequence": str(hit.getSequence()),
                "score": hit.getScore(),
                "charge": hit.getCharge(),
                "rt": pep_id.getRT(),
                "mz": pep_id.getMZ()
            })

df = pd.DataFrame(rows)
print(f"Identified peptides at 1% FDR: {len(df)}")
print(f"Unique sequences: {df['sequence'].nunique()}")
df.to_csv("peptide_identifications.csv", index=False)