mioXpektron 0.0.3

A comprehensive Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) data processing toolkit

mioXpektron

A comprehensive Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) data processing toolkit for advanced signal processing, peak detection, and calibration.

Features

mioXpektron provides a complete pipeline for ToF-SIMS data analysis:

Core Modules

• Baseline Correction - Multiple baseline correction algorithms including AirPLS, AsLS, and adaptive methods
• Denoising - Advanced noise filtering strategies: wavelet transforms, Gaussian filters, median filters, and Savitzky-Golay smoothing
• Peak Detection - Robust peak detection with automatic noise estimation and overlapping peak resolution
• Calibration - Flexible mass spectrum recalibration with multiple TOF models, explicit autodetect fallback policies, bootstrap or m/z-based channel detection, and optional two-pass reference-mass screening
• Normalization - 18 normalization strategies including TIC, SNV, robust SNV, selected-ion, multi-ion reference, PQN, mass-stratified PQN, Pareto, and automated evaluation/ranking
• Adaptive Parameterization - Opt-in data-driven estimation of pipeline thresholds (auto_tune=True) for calibration tolerance, outlier rejection, normalization targets, and more
• Visualization - Publication-ready plotting tools for spectra and peak analysis
• Batch Processing - High-throughput data processing utilities
• Pipeline - End-to-end automated processing pipeline

Installation

From PyPI

pip install mioXpektron

From Source

git clone https://github.com/kazilab/mioXpektron.git
cd mioXpektron
pip install -e .

Development Installation

pip install -e ".[dev]"

Quick Start

Basic Usage

import mioXpektron as mx

# Import your ToF-SIMS data (.txt with m/z + Intensity or .csv with mz + corrected_intensity/intensity)
mz, intensity, sample_name, group = mx.import_data("path/to/your/data.txt")

# Denoise the spectrum
denoised = mx.noise_filtering(intensity, x=mz, method='wavelet')

# Correct baseline
corrected = mx.baseline_correction(denoised, method='airpls')

# Detect peaks
peaks = mx.detect_peaks_with_area(
    mz_values=mz,
    intensities=corrected,
    sample_name=sample_name,
    group=group,
    min_snr=3.0,
    noise_model="mz_binned",
)

# Visualize results
plot = mx.PlotPeak(
    mz_values=mz,
    raw_intensities=intensity,
    sample_name=sample_name,
    corrected_intensities=corrected,
)
plot.plot()

Automated Pipeline

import glob
from mioXpektron import run_pipeline, PipelineConfig

# Configure the pipeline
config = PipelineConfig(
    denoise_method='wavelet',
    baseline_method='airpls',
    normalization_target=1e6,
)

# Run end-to-end processing (returns intensity and area DataFrames)
files = glob.glob("path/to/files/*.txt")
intensity_df, area_df = run_pipeline(files, config=config)

Adaptive Parameterization

Let the pipeline derive optimal parameters from your data instead of using fixed defaults:

from mioXpektron import FlexibleCalibrator, FlexibleCalibConfig

config = FlexibleCalibConfig(
    reference_masses=[1.0073, 27.0229, 29.0386, 41.0386, 57.0699, 104.1075],
    calibration_method="quad_sqrt",
    auto_tune=True,  # derives tolerance, screening, and breakpoints from data
)

calibrator = FlexibleCalibrator(config)
summary = calibrator.calibrate(files)

The pipeline also supports auto_tune:

config = PipelineConfig(auto_tune=True)  # derives mz_tolerance and normalization_target
intensity_df, area_df = run_pipeline(files, config=config)

See mioXpektron.adaptive for individual estimator functions.

Calibration

import glob
from mioXpektron import FlexibleCalibrator, FlexibleCalibConfig

# Flexible calibration with screened reference masses
config = FlexibleCalibConfig(
    reference_masses=[1.0073, 27.0229, 29.0386, 41.0386, 57.0699, 104.1075],
    calibration_method="quad_sqrt",
    autodetect_method="parabolic",
    autodetect_fallback_policy="max",
    autodetect_strategy="mz",
    auto_screen_reference_masses=True,
    output_folder="calibrated_spectra",
)

calibrator = FlexibleCalibrator(config)
files = glob.glob("path/to/spectra/*.txt")
summary_df = calibrator.calibrate(files)
print(calibrator.last_reference_masses_used)

Batch Processing

import glob
from mioXpektron import BatchDenoising, batch_tic_norm

# Batch denoising
files = glob.glob("path/to/files/*.txt")
batch_denoiser = BatchDenoising(files, method='wavelet')
batch_denoiser.run(output_root="output_files", folder_name="denoised_spectrums")

# Batch TIC normalization (accepts a glob pattern)
output_paths = batch_tic_norm("data/*.txt", output_dir="normalized_spectra")

Advanced Features

Denoising Method Selection

The recommended workflow is cohort-level selection with DenoisingMethods.compare_across_files(...). This evaluates peak preservation and denoising jointly, applies explicit pass/fail criteria, and by default excludes derivative filters from the search grid.

from pathlib import Path
from mioXpektron import DenoisingMethods
from mioXpektron.denoise.denoise_batch import load_txt_spectrum

files = sorted(Path("calibrated_spectra").glob("*_calibrated.txt"))

cohort_summary, per_file_summary, per_peak_detail = DenoisingMethods.compare_across_files(
    file_paths=files,
    min_mz=500.0,
    max_mz=520.0,
    include_derivatives=False,
    save_summary=True,
)

preview = load_txt_spectrum(files[0])
axis = preview.get("mz")
if axis is None or axis.size == 0:
    axis = preview.get("channel")

dm = DenoisingMethods(axis, preview["intensity"])
best_params = dm.method_parameters(
    cohort_summary,
    basis="constrained_pareto_then_snr",
    require_pass=False,
)

dm.plot(cohort_summary, top_k=3)
dm.denoise_check(best_params, sample_name=files[0].stem, mz_min=500.0, mz_max=520.0)

To tighten the scientific gates, pass selection_criteria={...} to compare(...), compare_in_windows(...), or compare_across_files(...).

Baseline Evaluation

Systematically evaluate baseline correction methods across a set of spectra:

from mioXpektron import BaselineMethodEvaluator

evaluator = BaselineMethodEvaluator(files=["data/*.txt"])
results_df = evaluator.evaluate()
evaluator.plot()

Normalization Evaluation

Evaluate and rank normalization methods on labelled spectra:

from mioXpektron import NormalizationEvaluator

evaluator = NormalizationEvaluator(
    files=["spectra/"],  # directory with .txt and/or baseline-corrected .csv spectra
    methods=["tic", "robust_snv", "pqn", "mass_stratified_pqn", "log"],
    method_kwargs_map={
        "mass_stratified_pqn": {
            "strata": [(0.0, 100.0), (100.0, 400.0), (400.0, float("inf"))],
        },
    },
)
results_df = evaluator.evaluate()
evaluator.plot()
evaluator.print_summary()

For cohort-level normalization on baseline-corrected CSV outputs, the repository notebook NoteBooks/_06_Normalization.ipynb builds a shared m/z grid, supports linear, pchip, akima, makima (SciPy >= 1.13), and cubic interpolation, ranks normalization methods, previews overlays, and exports the winning method. mass_stratified_pqn is included in the default notebook evaluation; multi_ion_reference is available when users provide multi_ion_reference_mz with optional multi_ion_reference_values.

Overlapping Peak Resolution

Detect and visualize overlapping peaks across a dataset:

from mioXpektron import check_overlapping_peaks2

check_overlapping_peaks2(
    data_dir="path/to/spectra",
    file_pattern="*.txt",
    mz_min=100.0,
    mz_max=200.0,
)

Documentation

For detailed documentation on each module:

• Adaptive Parameterization: See docs/modules/adaptive.rst
• Denoising: See denoise_doc.md
• Baseline Correction: See COLUMN_NAMING.md
• Calibration: See DEBUG_README.md
• Sphinx docs: See docs/modules/calibration.rst and docs/modules/detection.rst

Dependencies

• numpy >= 1.20.0
• pandas >= 1.3.0
• polars >= 0.18.0
• scipy >= 1.7.0
• matplotlib >= 3.4.0
• PyWavelets >= 1.1.0
• pybaselines >= 1.0.0
• scikit-learn >= 1.0.0
• joblib >= 1.0.0
• tqdm >= 4.60.0

Requirements

• Python 3.10 or higher

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Maintainer

• Developed by: Data Analysis Team @KaziLab.se
• Contact: [e-mail]
• Copyright: @kazilab.se

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use mioXpektron in your research, please cite:

@software{mioxpektron,
  author = {Data Analysis Team @KaziLab.se},
  title = {mioXpektron: A ToF-SIMS Data Processing Toolkit},
  year = {2026},
  url = {https://github.com/kazilab/mioXpektron}
}

Acknowledgments

mioXpektron builds upon established signal processing algorithms and the excellent scientific Python ecosystem. Parts of this documentation were created with assistance from ChatGPT Codex and Claude Code.

Support

For issues, questions, or contributions, please visit:

• Issues: https://github.com/kazilab/mioXpektron/issues
• Documentation: https://github.com/kazilab/mioXpektron#readme