Source: https://github.com/aipoch/medical-research-skills

When to Use

• Converting, validating, or batch-processing mass spectrometry files (e.g., mzML, mzXML, MGF) as part of a pipeline.
• Cleaning raw spectra before downstream analysis (smoothing, baseline correction, denoising, peak picking).
• Detecting and linking isotope patterns / features for proteomics or metabolomics feature tables.
• Running identification-oriented steps where peptide/protein identification integration is required.
• Building custom computational MS workflows in Python while leveraging OpenMS algorithms.

Key Features

• MS File I/O: Read/write common MS formats (mzML, mzXML, MGF).
• Signal Processing: Smoothing, baseline correction, filtering, and peak picking.
• Feature Detection: Isotope pattern detection and feature linking utilities.
• Identification Support: Hooks for peptide identification workflows via OpenMS-compatible components.
• Scripted Workflows: A ready-to-use “Load → Process → Analyze” workflow entry point.

Dependencies

Install the following Python packages:

• pyopenms (version: compatible with your OpenMS/PyOpenMS distribution)
• pandas (version: latest recommended)
• numpy (version: latest recommended)

Installation:

uv pip install pyopenms pandas numpy

Example Usage

A complete runnable example using the provided workflow script (scripts/process_ms.py):

# run_example.py
from scripts.process_ms import run_workflow

def main():
    # Load -> Process -> Analyze
    # The script is expected to read the input mzML and apply optional filtering.
    result = run_workflow("data.mzML", apply_filter=True)

    # The returned object depends on the implementation of run_workflow.
    # Common patterns include a processed experiment, a feature map, or a summary dict.
    print("Workflow finished.")
    print(result)

if __name__ == "__main__":
    main()

Run:

python run_example.py

For manual/custom workflows, see:

• File operations: references/file_io.md
• Signal processing algorithms: references/signal_processing.md

Implementation Details

• Binding Layer: This skill uses PyOpenMS, the Python bindings for the OpenMS C++ library, to expose core computational MS algorithms.
• Workflow Pattern: The default script follows a standard pipeline structure:
  1. Load an MS run from disk (e.g., mzML).
  2. Process spectra (optional filtering/smoothing/baseline correction).
  3. Analyze results (e.g., peak picking, feature detection, or downstream summaries).
• Configurable Processing: The apply_filter flag in run_workflow(...) is intended to toggle one or more preprocessing steps; exact filters and parameters should be documented in scripts/process_ms.py and the referenced guides.
• Algorithm Reference: Detailed descriptions of available filters and peak pickers, including parameterization, are maintained in references/signal_processing.md.