#39 · Metabolomics

Learned and benchmarked computational tools for MS metabolomics feature processing

The shared thread is computational processing of mass spectrometry data, with each item addressing a different stage. Bromo, a transformer model, ranks peptide precursors by relative MS response with charge state for targeted experiments, outperforming sequence-based methods and allowing fine-tuning. maxiM/Ze applies OpenCV image recognition to LC-HRMS extracted ion chromatograms for automated peak alignment and reports validation. TIMS-Bench shifts to evaluation, benchmarking MZmine, MS-DIAL, and MetaboScape across ten public and three benchmark datasets while releasing ground-truth data and code. The contribution is methodological rather than a unified result, and all three items are preprints that have not undergone peer review.