MARLIN: De Novo Molecular Structure Elucidation from Tandem Mass Spectra without a Ground-Truth Formula
Not on the map. Why?
Abstract (arXiv)
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.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 13 of its 30 references that are in the map. First posted on arXiv 6 Jul 2026.
Builds on
13 of 30 references.
References · 13
- 2017Attention Is All You Needcited by 59,932
- 2020Denoising Diffusion Probabilistic Modelscited by 20,851
- 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,665
- 2024Simple and Effective Masked Diffusion Language Modelscited by 800
- 2022DiGress: Discrete Denoising diffusion for graph generationcited by 503
- 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Modelscited by 385
- 2023Gotta be SAFE: A New Framework for Molecular Designcited by 45
- 2025DiffMS: Diffusion Generation of Molecules Conditioned on Mass Spectracited by 28
- 2025MADGEN: Mass-Spec attends to De Novo Molecular generationcited by 16
- 2025MS-BART: Unified Modeling of Mass Spectra and Molecules for Structure Elucidationcited by 9
- 2026Comparative Analysis of Formula and Structure Prediction from Tandem Mass Spectracited by 3
- 2026De Novo Molecular Generation from Mass Spectra via Many-Body Enhanced Diffusioncited by 3
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026MS-GPT: Rethinking MS/MS De Novo Structure Elucidation as Spectrum-Induced Posterior Querying of a Molecule-Language Model
- 2026De novo molecular structure elucidation from mass spectra via flow matching
- 2026Explainable Molecular Structure Inference from GC--MS with Diffusion Models and LLM Reranking
- 2025Atomic Diffusion Models for Small Molecule Structure Elucidation from NMR Spectra
- 2026How well can off-the-shelf LLMs elucidate molecular structures from mass spectra using chain-of-thought reasoning?
- 2026GLACIER: Rethinking Mass Spectrum Prediction as an Object Detection Problem
- 2026FlowMS: Flow Matching for De Novo Structure Elucidation from Mass Spectra
- 2025DiSE: A diffusion probabilistic model for automatic structure elucidation of organic compounds