DASMixer v0.7.3

v0.7.3

Date: 2026-09-28

---

New Features

• MaxQuant project import — import MaxQuant results directly from mqpar.xml + .apl files, with GUI wizard (parse → configure → progress) and CLI command dasmixer-cli import maxquant. Supports optional FASTA protein import, contaminant filtering, and duplicate handling.
• MGF Export: Merge MGF modes — choose between one merged file, per-sample files, or per-spectre-file grouping. New Add timestamp and Replace SCANS with Spectre ID options for reproducible external processing.
• MGF Export: batch performance — single-query get_full_spectrum_for_export replaces per-spectrum DB round-trips. New Sequence contains and File Suffix filters for selective export.
• PCA Report — Protein mode — new point grouping where each point is a (protein, subset) pair with LFQ profiles sorted by abundance. Top-N proteins selected by variance, with ROC/AUC per subset.
• Importable FDR, e-value, q-value — three new optional columns in Identification table imported from source files via ColumnRenames (persist through Clear Calculations).
• Max FDR filter — new threshold in Preferred identification tool settings (max_fdr, default 0.01) and a table filter on the Peptides tab, enforcing fdr <= threshold during preferred selection and peptide browsing.
• Peptides table — FDR, e-value, q-value columns — new read-only columns (hidden by default, configurable via column selector).
• User-configurable PTM definitions — PTM data moved from hardcoded constants to editable CSV files in the configuration folder (~/.config/dasmixer/ptm/), generated from built-in defaults on first run.
• Open Configuration Folder button — quick access to the configuration directory from the Settings screen, opening ~/AppData/dasmixer/ (Windows) or ~/.config/dasmixer/ (Linux/macOS) in the native file manager.

Bug Fixes

• Identification import with multiple spectral files — identification files are now matched to spectra files by basename (positional pairing) instead of always picking the first one, preventing silent data corruption when a sample has several MGF sources.
• Project import: missing protein taxonomy — importing a project no longer drops taxon_id and organism_name from proteins (regression since v0.2.0).
• Project import: missing identification metrics — importing a project no longer loses quality, lcrr, has_ptm, override_pepmass, and unconfirmed_ptms from identifications (regression since v0.7.1).
• Project import: version guard — merging projects with different PROJECT_VERSION now raises a clear error with upgrade instructions instead of silently producing incomplete rows.
• iBAQ NaN intensity propagation — spectra with NULL intensity in the DB produced NaN values that poisoned sum() in calculate_ibaq_value and normalize_values, causing ~95 % of iBAQ results to be saved as NULL. Fixed by filtering NaN/None from intensities before summation and replacing them with 0.0 in normalization. Added diagnostic logging (logger.exception/logger.warning) throughout the LFQ calculation pipeline.