MARLIN: De Novo Molecular Structure Elucidation from Tandem Mass Spectra without a Ground-Truth Formula

2026-07-06

AI Summary

In untargeted tandem mass spectrometry, the identification of numerous novel small molecules remains challenging due to the absence of reference spectral libraries and the reliance of existing de novo interpretation methods on known molecular formulas. To address this, this work proposes MARLIN—the first end-to-end structure elucidation method that operates without requiring a molecular formula as input. MARLIN employs a self-supervised encoder to predict molecular fingerprints, coupled with a chunked diffusion language model to generate candidate structures, while incorporating provably safe mass-shell constraints to guarantee exact agreement between generated structures and the observed precursor mass. Additionally, it introduces symmetric noise objectives and diversity-aware sampling to enhance robustness and structural coverage. On the NPLIB1 benchmark, MARLIN significantly outperforms existing approaches in exact match rate, structural distance, and fingerprint similarity under molecular-formula-free conditions, and recovers molecular formulas with accuracy approaching that of specialized predictors.

Abstract

Untargeted tandem mass spectrometry (MS/MS) detects thousands of small molecules per biological sample, yet most go unidentified because they are absent from spectral libraries. These uncharacterized metabolites and natural products are precisely the compounds that matter for drug discovery, biomarker research, and exposomics. Computational de novo structure elucidation could close this gap, but almost all state-of-the-art methods assume the ground-truth molecular formula is known, an oracle that does not exist for genuinely novel compounds and is itself predicted with substantial error. We present MARLIN, a de novo method that elucidates structures directly from a spectrum with no molecular formula at any stage. A self-supervised encoder predicts a molecular fingerprint from the raw peaks, and a block-diffusion language model generates candidate structures conditioned only on the fingerprint and the instrument-measured precursor mass. A provably safe mass-shell constraint keeps every candidate consistent with the measured mass without fixing the atom inventory, and candidates are accepted by exact parts-per-million mass agreement. A symmetric noise objective absorbs encoder error, and a candidate-diversity mechanism keeps the candidates from collapsing to a single structure. On the NPLIB1 benchmark, MARLIN is the strongest method evaluated without a ground-truth formula across exact-match accuracy, structural distance, and fingerprint similarity, and it recovers the correct molecular formula as a byproduct about as often as a dedicated predictor without ever using one. MARLIN enables reliable de novo structure elucidation in the realistic discovery regime where the molecular formula is unavailable.

Problem

• de novo structure elucidation
• tandem mass spectrometry
• molecular formula
• untargeted metabolomics
• spectral library

Innovation

• de novo structure elucidation
• tandem mass spectrometry
• molecular formula-free
• diffusion language model
• mass-shell constraint