MSDIAL-v5.5.261018 console

Summary

This release adds internal-standard normalization and method-file templates to the console application, introduces automatic alignment-only retention time correction for LC-MS, and makes annotation results more traceable across exports. It also fixes many method-file parsing issues and crashes in the console workflow.

Feature Updates

• Added the normalize command for internal-standard normalization of aligned results, including the option to specify internal standards by alignment ID. #795, #830 by @htsugawa
• Added the template command, which writes a method file holding the built-in defaults for each processing mode. Relative paths in method files are now resolved the same way in every mode. #819 by @YukiMatsuzawa
• Added automatic alignment-only retention time correction for LC-MS, and improved anchor selection by judging anchors against their local neighbors. #810, #826 by @htsugawa
• Added alignment provenance exports (AlignResult-*.mdpeakid.tsv and the optional AlignResult-*.mdprovenance.tsv) that trace each alignment spot back to its source peaks. #791 by @htsugawa
• Exported every threshold-passing annotation candidate, and reported ambiguous annotations as "A or B" in mzTab-M. #793 by @htsugawa
• Recorded the evidence behind each annotation (what was compared, the resulting score, the library used, and the MS-DIAL build), and made every output report it consistently. #800 by @htsugawa
• Added tiered LC-MS annotation settings, allowing each MSP annotator to select its own target omics and LBM priority. #783 by @htsugawa
• Added a lipid retention-time-order check that assigns "Confirmed" to lipid annotations following the reversed-phase elution rules. #832 by @htsugawa
• Supported AIF data with collision energies read from mzML activation, and multi-collision-energy AIF alignment. #825 by @htsugawa

Improvements

• Summarized verbose method-key warnings, and recorded the keys read by LC-MS processing in the method-key record (<method>.keys.json). #818, #811 by @htsugawa
• Suppressed redundant raw-data reading logs during LC-MS alignment. #777 by @htsugawa
• Recognized the MS-DIAL 5 suggested-annotation name prefixes (no MS2: and low score:). #784 by @htsugawa
• Recorded each raw data file's format at registration so that the mzTab-M ms_run-format is kept even when the raw data is unavailable. #835 by @htsugawa
• Improved annotation and processing performance, including spectral entropy similarity (about 25 times faster), MS2 peak extraction, spectrum peak shrinking, imaging spectrum binning, molecular networking, and peak group traversal. #827, #775, #780, #792, #812, #821 by @YukiMatsuzawa
• Built the Linux and macOS console archives with the vendor-unsupported configuration. by @YukiMatsuzawa

Bug Fixes

• Fixed blank filtering threshold parsing in the console, and made Blank filtering mode and fold change the canonical blank filtering settings. #773 by @Fraximov, #823 by @YukiMatsuzawa
• Fixed several method-file keys that had no effect or were applied incorrectly:
  • "Searched adduct ions" now takes effect. #815 by @YukiMatsuzawa
  • On/off keys are now read in both directions. #813 by @YukiMatsuzawa
  • "Use CCS for LBM-based annotation filtering" now sets the LBM parameter. #816 by @YukiMatsuzawa
  • The RT correction library is loaded once from the final path, and duplicated keys are last-wins everywhere. #817 by @YukiMatsuzawa
• Fixed a crash in the alignment export of lcimms mode. #838 by @YukiMatsuzawa
• Fixed the DIMS LBM annotator to use the LBM search parameters. #814 by @YukiMatsuzawa
• Fixed AIF processing failures and the nearest-apex seed scan selection in AIF deconvolution. #825 by @htsugawa
• Fixed processing that stopped at a Y/N prompt when the input mixed raw data formats. #831 by @htsugawa
• Fixed GC-MS Fiehn RI compatibility in the console, and applied the GC-MS RI tolerance on the selected retention index scale. #779, #802 by @htsugawa
• Fixed a KeyNotFoundException when exporting alignment matrices in alignment-light mode. #809 by @htsugawa
• Fixed annotation score exports: uncomputed scores are reported as null, the MS/MS term is skipped when no spectrum was compared, and sentinel values are no longer published as measurements. #785, #789, #794 by @htsugawa
• Fixed a crash when loading projects with large libraries. #799 by @YukiMatsuzawa
• Fixed missing defaults for automatic RT correction parameters in older projects. #837 by @htsugawa
• Fixed ether LPE annotation and MS/MS characterization. #803, #808 by @mikikot113
• Removed the zero-intensity C3+H peak from plasmalogen EAD/EID spectra. #828 by @YukiMatsuzawa

Changes That Affect Results

Reanalyzing data with the same method file may give different results from the previous console release:

• Adducts listed in "Searched adduct ions" are now searched; previously only [M+H]+ or [M-H]- was searched. #815
• The suggested compound may change for precursor-only annotations. #789
• The GC-MS RI tolerance is now interpreted on the selected retention index scale. #802

New Contributors

• @Fraximov (François-Xavier Lehr, EMBL Heidelberg) made their first contribution in #773. Thank you!

Full Changelog: MSDIAL-v5.5.260820...MSDIAL-v5.5.261018