matchms

Spectral matching identifies unknowns by scoring experimental spectra against a reference library. You can compute similarity using CosineGreedy, ModifiedCosine, or NeutralLossesCosine to annotate chemical information and rank potential matches.

Yes, you can import MGF, MSP, and mzML files. Metadata harmonization applies default filters, normalization, and peak filtering to ensure consistent spectral comparisons across different file formats in your metabolomics pipeline.

CosineGreedy computes standard cosine similarity, while ModifiedCosine allows matching peaks with shifted precursor masses. NeutralLossesCosine adds neutral loss comparisons, and structural fingerprinting can further improve metabolite ranking accuracy.

Yes, mzML is supported alongside MGF and MSP. The processing pipeline harmonizes metadata from these formats to facilitate end-to-end workflows including library matching, quality control, and reproducible metabolite annotation.

Apply a standard processing pipeline with default_filters, normalization, peak filtering, and a chosen similarity function. Optional fingerprinting and validation steps can be added to ensure reproducible library curation and annotation results.

NeutralLossesCosine is useful when comparing spectra where neutral losses carry diagnostic structural information. It complements CosineGreedy and ModifiedCosine by scoring similarity based on fragmentation patterns beyond direct peak matching.