Simulation Evaluation Data Capture

simulation-evaluation-data-capture

Summary

Capture evaluation metrics and write simulated MS/MS scans to mzML format during a virtual mass spectrometry acquisition loop. This skill bridges in-silico fragmentation strategy prototyping with downstream comparative analysis by preserving both scan results and performance metadata.

When to use

When you have simulated DDA (data-dependent acquisition) scans from a ViMMS Environment and need to (1) quantify how well the simulated acquisition matched real or reference data (via evaluation metrics), and (2) export the results as standards-compliant mzML files for comparison with empirical LC-MS/MS data or alternative fragmentation strategies.

When NOT to use

• Input is only a single chemical spectrum or a static database query—this skill requires a full time-resolved acquisition loop through an Environment.
• No reference or comparison target exists—evaluation metrics are most valuable when contrasted against real or alternative strategy results.
• mzML export is not supported by your ViMMS build or you lack write permissions to the output path.

Inputs

• Environment object with configured mass spectrometer and controller
• Simulated scan list (generated during env.run())
• Controller fragmentation strategy parameters (N for TopN, isolation window m/z, intensity thresholds)

Outputs

• mzML file containing simulated MS1 and MS/MS scans
• Evaluation metrics object (scan counts, chemical coverage, intensity statistics)
• Downstream comparison data for strategy validation

How to apply

During Environment instantiation, set save_eval=True to enable automatic capture of evaluation metrics (e.g., scan count, chemical coverage, intensity distribution statistics) as the acquisition loop runs. After env.run() completes, invoke Environment.write_mzML() to serialize the simulated scan list to mzML format. The evaluation metrics are computed scan-by-scan as the TopNController or other controller makes acquisition decisions, allowing you to assess strategy efficacy (e.g., how many of the top-intensity chemicals were selected, retention time accuracy) without manual post-processing. Export the resulting mzML alongside evaluation summaries to enable blind or side-by-side comparison with reference data, such as real Beer1pos mzML samples.

Related tools

• ViMMS (Framework providing Environment class, scan simulation loop, and write_mzML() method for metric capture and file export) — https://github.com/glasgowcompbio/vimms
• TopNController (Fragmentation strategy controller that determines which precursor ions are selected for MS/MS; its decisions are tracked in evaluation metrics) — https://github.com/glasgowcompbio/vimms
• IndependentMassSpectrometer (Virtual mass spectrometer that generates MS1 and MS/MS scans fed into the evaluation loop) — https://github.com/glasgowcompbio/vimms
• ChemicalMixtureFromMZML (Extracts regions of interest (ROIs) from empirical mzML data as UnknownChemical objects for use as input to the virtual spectrometer and evaluation baseline) — https://github.com/glasgowcompbio/vimms

Examples

env = Environment(mass_spectrometer=ms, controller=TopNController(N=10, isolation_width=1.0), retention_time_range=(0, 1E5), save_eval=True); env.run(); env.write_mzML('simulated_dda.mzML')

Evaluation signals

• mzML file is valid and parseable by standard MS data tools (e.g., Proteowizard, mzmine); check file header and scan index integrity.
• Evaluation metrics are non-null and contain reasonable scan counts (>0) and intensity ranges matching the input chemical mixture.
• Number of MS/MS scans generated equals or is less than the TopN parameter × number of MS1 scans, reflecting correct acquisition logic.
• Retention time values in output mzML span the specified acquisition window (start_rt to stop_rt); no scans outside this range.
• Simulated mzML scan statistics (e.g., median precursor m/z, median intensity) are comparable in order of magnitude to the reference Beer1pos data.

Limitations

• Evaluation metrics depend on correct peak picking via MZMine parameters in PeakPicking.py; suboptimal or mismatched peak detection settings will distort coverage and intensity statistics.
• mzML export does not preserve all ViMMS internal metadata (e.g., controller decision history, intermediate chemical states); downstream tools may need custom parsing for full audit trails.
• Real-time evaluation during env.run() can add computational overhead; very large chemical mixtures (>10,000 compounds) or long RT windows may slow metric collection.
• Evaluation is relative to the input chemical list; if the chemical extraction (via ChemicalMixtureFromMZML) misses low-abundance species or has poor ROI filtering, evaluation metrics may underestimate true strategy performance.

Evidence

• [methods] Create an Environment with the mass spectrometer and controller, set the retention time range to the total acquisition window, and enable save_eval=True to capture evaluation metrics.: "Create an Environment with the mass spectrometer and controller, set the retention time range to the total acquisition window, and enable save_eval=True to capture evaluation metrics."
• [methods] Execute env.run() to simulate the DDA acquisition loop.: "Execute env.run() to simulate the DDA acquisition loop."
• [methods] Write the simulated scans to mzML format using Environment.write_mzML() for downstream comparison.: "Write the simulated scans to mzML format using Environment.write_mzML() for downstream comparison."
• [readme] a flexible and modular framework designed to simulate fragmentation strategies in tandem mass spectrometry-based metabolomics: "a flexible and modular framework designed to simulate fragmentation strategies in tandem mass spectrometry-based metabolomics"
• [readme] You can evaluate diverse fragmentation strategies using real data, and extract the scan results as mzML files.: "You can evaluate diverse fragmentation strategies using real data, and extract the scan results as mzML files."
• [other] The evaluation helpers rely on peak picking using MZMine parameters defined in PeakPicking.py.: "The evaluation helpers rely on peak picking using MZMine parameters defined in PeakPicking.py."