MS-DIAL 4.9.22118: Peak list export "deconvoluted" exports centroid data instead of true deconvoluted spectra for MSE single-sample data

[Lzlulu]

Dear MS-DIAL developers,

I am using MS-DIAL 4.9.22118 to analyze Waters MSE (DIA) data from a single traditional Chinese medicine formula sample. I encountered an issue when trying to export deconvoluted MS2 spectra for GNPS molecular networking.

Problem description:

1. MS2Dec deconvolution works correctly in the GUI - I can see clean "Deconvoluted MS/MS spectrum" in the Peak spot window (e.g., only 2-3 major fragments like m/z 113.9574 and 149.0176 for precursor 158.956).
2. However, when using File → Export → Peak list result → Spectra type: "deconvoluted", the exported .mgf file contains continuous ion series (e.g., m/z 60-159 for precursor 158.956), which appears to be centroid data rather than truly deconvoluted spectra.
3. The "centroid" option exports m/z 50-1000, which is clearly raw full-scan data.
4. The difference between "deconvoluted" and "centroid" exports is minimal - both contain many more fragments than what is shown in the GUI's Deconvoluted MS/MS spectrum view.

Data and settings:

• Data type: Waters Synapt G2-Si MSE (positive mode)
• Sample: Single sample (no alignment performed)
• MS-DIAL version: 4.9.22118
• Import mode: SWATH-MS + experiment file
• MS2Dec parameters: Sigma window 0.5, MS/MS abundance cut off 0, Exclude after precursor ion checked

What I've tried:

1. Waters2mzML conversion → same issue
2. Direct MS-DIAL import with SWATH-MS + experiment file → MS2Dec works in GUI but export fails
3. Different export options (Molecular spectrum networking export, Peak list export) → all export raw/centroid data, not deconvoluted
4. Checked .dcl and .pai2 files in project folder → binary format, cannot be read by external tools

Expected behavior:
The "deconvoluted" export option should produce .mgf files containing only the clean, deconvoluted fragments shown in the GUI's "Deconvoluted MS/MS spectrum" view (e.g., 2-10 major fragments per compound, not continuous ion series).

Questions:

1. Is this a known bug in version 4.9.22118?
2. Is single-sample MSE data fully supported for deconvoluted spectrum export?
3. Are there any workarounds to export the truly deconvoluted spectra?
4. What is the format of .dcl and .pai2 files? Is there any API or tool to extract deconvoluted spectra from them?

Thank you for your excellent work on MS-DIAL. I would greatly appreciate any guidance on this issue.

Best regards,
LZlulu