Skip to content

MS-GPT: Rethinking MS/MS De Novo Structure Elucidation as Spectrum-Induced Posterior Querying of a Molecule-Language Model

Jul 2026 · arXiv.org · Vol abs/2607.23607 · 0 citations · 43 references
Computer Science

TL;DR

MS-GPT is introduced, which recasts fingerprint-mediated de novo elucidation as spectrum-induced posterior querying of a conditional molecule-language model, and shows that efficient autoregressive molecular generation continues to improve recall with a little additional inference cost.

Abstract

Molecular structure elucidation from tandem mass spectra (MS/MS) is a central inverse problem in analytical chemistry. Most existing approaches to MS/MS identification remain tied to reference libraries or predefined candidate sets, whereas de novo methods aim to generate structures directly from spectra. A common de novo route predicts a molecular fingerprint from the spectrum and then decodes structures from it, enabling decoder pretraining on large molecule-only corpora. However, this paradigm creates a training-inference mismatch: the decoder is trained on oracle fingerprints computed from molecules, but at inference it is queried with a noisy spectrum-induced fingerprint posterior that is typically collapsed to a single thresholded fingerprint. We introduce MS-GPT, which recasts fingerprint-mediated de novo elucidation as spectrum-induced posterior querying of a conditional molecule-language model. MS-GPT conditions a molecule-language model on fingerprints and formulas, then converts the spectrum-induced posterior into a band of fingerprint queries near the oracle-fingerprint manifold through active-bit density calibration. Candidates sampled across this band are pooled and ranked by generation-frequency consensus. A lightweight LoRA adapter further mitigates domain-specific posterior bias while preserving the pretrained molecular prior. On NPLIB1 and MassSpecGym, MS-GPT sets a new state of the art, reaching Top-1/Top-10 exact-match accuracy of 29.8\%/41.1\% and 23.9\%/28.7\%, respectively. Candidate-pool scaling shows that efficient autoregressive molecular generation continues to improve recall with a little additional inference cost. The source code and model checkpoints are available at https://github.com/VIKI623/MS-GPT.

View source

Similar papers

Open access Sep 2026

Learning from tandem mass spectra at scale with a self-supervised foundation model for proteomics

InstaNovo-FM is presented, a self-supervised foundation model for bottom-up proteomics trained to reconstruct masked regions of tandem mass spectra and establishes a unified representation space for peptide fragmentation spectra, enabling robust transferability across the proteomics ecosystem.

M. Nieuwoudt, Marco Reverenna, Divanisha Patel et al. · 1 citation · ⚡1
Sep 2026

Language models as QSPR predictors: Unleashing potential through in-context learning and instruction tuning

The accuracy, generalizability, data scalability, and ease of use of this work demonstrate that FuelProp-LM can serve as a practical and flexible tool for fuel property prediction and play a significant role in future AI-assisted fuel design.

Chen-Yue Tao, Cheng-Cheng Liu, Chen-Xuan Li et al. · 0 citations
Open access Sep 2026

Genome-wide-scale prediction of compound−protein interactions using foundation and language models based on three-dimensional structures of compounds and proteins

DESTIN (dual structure-guided estimation of substance-target interaction networks), an ML framework for CPI prediction on a genome-wide scale from the three-dimensional structures of compounds and proteins exhibited good predictive performance, even for structurally complex and pharmacologically important membrane prot...

Yuga Moriyama, Yuki Matsukiyo, T. Kimura et al. · 0 citations
Sep 2026

TC-TopoRT: Tautomer-consistent and topology-aware retention time prediction for LC-MS metabolite annotation.

Accurate retention time (RT) prediction provides chromatographic evidence complementary to precursor mass and MS/MS spectra in LC-MS metabolomics. Such evidence is particularly valuable for metabolite annotation, where multiple structurally plausible candidates may exhibit similar spectral signatures. Here, we present...

Runquang Lai, Zhan-Qiang Liu, Bing-Yi Wang et al. · 0 citations
#machine learning Preprint Aug 2026

S3C-LLM: Skill-Code Guided Agentic Language Models for Spectrum-to-Structure Elucidation

S3C-LLM is introduced, a skill-guided and code-grounded agentic LLM for spectrum-to-structure elucidation that consistently outperforms current general LLMs and spectrum-specific models across spectra, while using less than 1/10th of SpectraLLM's training corpus.

Xuan-Le Zhao, Xinyu Cai, Xiang Cheng et al. · 0 citations
Open access Aug 2026

Charting the small-molecule universe from mass spectra with neuro-symbolic AI

AIMe (AI Molecule Explorer), a multi-agent neuro-symbolic AI framework that transforms the interpretation of unknown spectra into an omics-scale exploration across the known structural space, providing chemically interpretable annotations, is introduced.

U. U. Acikalin, Dieqiao Feng, Aaron M. Ferber et al. · 0 citations

Related blog posts

MIT News · Artificial Intelligence Nov 17, 2026

Learning the language of machine learning

By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.