MetaPhoenix 1.1.0

MetaPhoenix 1.1.0 is a maintenance release of the software published with Ion entropy and accurate entropy-based FDR estimation in metabolomics (Briefings in Bioinformatics, 2024). The project was previously named MSLibrary; the original code is archived on the master branch as v1.0.0.

The decoy generation and scoring methods are unchanged. Databases created with MSLibrary, including the one used for the paper, can be read without migration.

Download and run

Requires Java 17 or newer and MongoDB.

java -jar metaphoenix-core-1.1.0.jar --command.runner.enabled=true

See the README for the full workflow.

What's new

Command line

• Import GNPS (-t gnps), MassBank-Europe (-t massbank) and MoNA (-t mona) libraries directly with import.
• Reproducible decoy libraries with decoy -seed.
• Running decoy again no longer adds a second decoy set to an existing decoy library, which doubled the decoy hits and the estimated FDR. An existing decoy library is kept unless -overwrite is given.
• Clear messages for missing or invalid parameters; a failing command no longer stops the command runner.

Fixes

• identify reported every target hit as FDR-controlled when no score threshold reached the requested FDR. It now reports no hits and a warning.
• MSP, MassBank and MoNA parsers kept only the first word of text fields such as compound names, comments and instruments.
• MassBank and MoNA parsers could store the InChIKey as the InChI and the instrument type as the instrument.
• MGF parser dropped the first peak of every spectrum.
• Entropy was NaN when an intensity was zero, and the best hit was lost when its score was 0.
• The MongoDB connection setting was ignored after the Spring Boot upgrade.

Performance

• Faster ion entropy–based decoy generation (binary search for precursor and fragment windows).

Maintenance

• Spring Boot 4.1 and updated dependencies; tested on Java 17, 21 and 25 in CI.
• Renamed from MSLibrary to MetaPhoenix (Maven modules metaphoenix-client and metaphoenix-core, package net.csibio.metaphoenix).
• Removed unused proteomics, peptide and DIA code.
• Unit tests for the parsers, decoy generation, similarity and the command line.
• Documentation: rewritten README, differences between the paper text and the code, and CITATION.cff.

Citation

An S, Lu M, Wang R, Wang J, Jiang H, Xie C, Tong J, Yu C. Ion entropy and accurate entropy-based FDR estimation in metabolomics. Briefings in Bioinformatics. 2024;25(2):bbae056. doi:10.1093/bib/bbae056