Mzml Format Output Generation And Validation

mzML-Format Output Generation and Validation

Summary

Export simulated LC-MS/MS scan data from ViMMS to industry-standard mzML format, enabling downstream analysis, archival, and comparative evaluation against real or baseline acquisition strategies. This skill bridges virtual simulation output to standard proteomics/metabolomics workflows.

When to use

After running a ViMMS simulation loop with a fragmentation controller (e.g., TopNController, WeightedDEWController) to completion, you need to serialize the in-memory scan list to a portable, standards-compliant file format that can be processed by third-party tools (OpenMS, MZmine) or compared against reference acquisitions.

When NOT to use

• Input scans have not yet been generated (env.run() has not completed) — wait for simulation loop to finish.
• Output format is required by a tool that demands NetCDF or proprietary binary formats (e.g., vendor-specific .raw files) — mzML is text-based and may be slower to parse for very large datasets.
• Evaluation metrics are needed before mzML export — prioritize EvaluationData pickle capture instead, as mzML is primarily for archival and secondary analysis.

Inputs

• Environment object (instantiated with IndependentMassSpectrometer, controller instance, and save_eval=True)
• In-memory scan list from completed env.run()
• Specification of output file path (out_file) and directory (out_dir)

Outputs

• mzML file (XML-formatted mass spectrometry data)
• EvaluationData pickle (optional; contains coverage, intensity, and fragmentation statistics)
• Metadata: scan count, retention time range, m/z range, intensity distribution

How to apply

Within the ViMMS workflow, after instantiating an Environment with a mass spectrometer and controller and calling env.run() to accumulate scans, invoke env.write_mzML() to serialize the complete scan list to mzML format. Set the save_eval=True flag when creating the Environment to enable simultaneous capture of EvaluationData pickles. The output mzML file contains all MS1 and MS/MS spectra in their acquisition order, preserving scan metadata (retention time, m/z, intensity, isolation windows, fragmentation parameters). Validate the mzML output by opening it in peak-picking tools (e.g., OpenMS) or by loading the parallel EvaluationData pickle to cross-check scan counts, intensity ranges (e.g., cumulative intensity thresholds), and fragmentation coverage metrics against baseline controller results.

Related tools

• VIMMS (Core simulation framework; provides Environment class and write_mzML() method to export scan data) — https://github.com/glasgowcompbio/vimms
• OpenMS (Processes mzML output from simulation to compute fragmentation coverage metrics and validate file integrity)
• Python (Language for executing Environment.write_mzML() and post-hoc mzML parsing)

Examples

from vimms.Env import Environment
from vimms.Controller import TopNController
from vimms.MassSpec import IndependentMassSpectrometer

env = Environment(ms, controller, min_time=0, max_time=1440, save_eval=True, out_file='simulation_output.mzML', out_dir='./results')
env.run()
env.write_mzML()

Evaluation signals

• mzML file exists at specified out_file path and is valid XML (parseable by standard XML tools and mzML validators).
• Scan count in mzML header matches the length of the in-memory scan list from env.run().
• EvaluationData pickle cumulative intensity and times_fragmented_summary values are consistent with MS1 and MS/MS peak intensities in the mzML file.
• Retention time, m/z, and isolation window values in mzML scans match the chemical mixture and controller parameters (e.g., isolation_width=1, rt_tol=15).
• mzML file can be successfully imported by downstream tools (OpenMS, MZmine) without parsing errors; peak picking yields expected coverage metrics (e.g., fragmentation rate within ±10% of baseline TopN controller).

Limitations

• mzML is a verbose, text-based format; exporting very large simulations (>10,000 scans) may incur significant I/O overhead and disk space.
• mzML export does not perform real-time validation; file correctness depends on the integrity of the Environment's scan list. Corrupted or incomplete simulations will propagate to mzML.
• mzML output lacks vendor-specific metadata (e.g., instrument tuning parameters, calibration curves) present in raw binary formats, limiting comparison to real instrument data beyond basic m/z, intensity, and retention time.
• Peak picking for downstream evaluation relies on external tools (OpenMS, MZmine) with their own parameter sensitivity; mzML alone does not guarantee reproducible coverage metrics without consistent peak-picking configuration.

Evidence

• [other] The Environment class provides write_mzML to export the generated scans: "The Environment class provides write_mzML to export the generated scans"
• [other] When running an Environment you can enable the save_eval flag: "When running an Environment you can enable the save_eval flag"
• [readme] ViMMS can extract scan results as mzML files: "extract the scan results as mzML files"
• [other] Call env.write_mzML() to serialize the scan list to mzML format: "Call env.write_mzML() to serialize the scan list to mzML format"
• [other] Processes mzML output from a simulation to compute fragmentation coverage: "Processes mzML output from a simulation (or real acquisition) to compute fragmentation coverage using OpenMS"