MassFlow

MassFlow is a lightweight Python toolkit for processing, cleaning, and analyzing tandem mass spectrometry (MS/MS) data. It leverages the matchms ecosystem to provide efficient spectral data handling and similarity calculations.

Features

• Spectral Cleaning: Automated metadata repair, peak filtering, and normalization.
• Format Conversion: Convert between MGF, MSP, JSON, and Pickle formats.
• Similarity Search: Calculate Cosine and Modified Cosine similarity scores between spectra.
• CLI & Library: Use as a command-line tool or import as a Python library.

Installation

Prerequisites

• Python 3.10+

Install from Source

git clone https://github.com/yourusername/MassFlow.git
cd MassFlow
pip install .

Development Setup

pip install -r requirements.txt

Usage

MassFlow provides a CLI entry point massflow (or MassFlow depending on installation).

Command Line Interface (CLI)

1. Clean and Convert a Library

Process an MGF or MSP file to apply filters (intensity, mz range, metadata cleanup) and save it in a new format.

# Clean an MSP file and save as Pickle (default)
massflow clean --input data/library.msp --output-dir processed_data/

# Clean an MGF file and save as MSP
massflow clean --input data/query.mgf --output-dir processed_data/ --format msp

Options:

• --input: Path to input library (.mgf or .msp).
• --output-dir: Directory to save the output.
• --format: Output format (pickle, msp, mgf, json). Default: pickle.

2. Similarity Search

(Functionality available in library, CLI wrapper coming soon)

massflow search ...

Python Library

You can use MassFlow modules directly in your Python scripts.

Processing Spectra

from MassFlow import processing, io

# Load and clean a library
spectra = processing.clean_mgf_library("data/test.mgf")

# The cleaning pipeline automatically applies:
# - Metadata repair (InChI/SMILES)
# - Peak filtering (intensity > 0.01, relative > 0.08)
# - Normalization

Similarity Calculations

from MassFlow import similarity

# Calculate Cosine scores
scores = similarity.calculate_cosscores(reference_spectra, query_spectra)

# Get top matches
matches = scores.scores_by_query(query_spectra[0], sort=True)
print(matches[:5])

Testing

Run the test suite using pytest:

pytest

Development

• Style: Codebase follows PEP8.
• Structure:
  • MassFlow/cli.py: CLI entry point.
  • MassFlow/processing.py: Core cleaning logic using matchms.
  • MassFlow/io.py: Input/Output handlers.
  • MassFlow/similarity.py: Similarity scoring functions.

License

MIT License