aiqbench analysis
MS-GPT advances de novo molecular structure elucidation by reframing the inverse problem as a spectrum-induced posterior query within a molecule-language model framework. This approach resolves training-inference mismatches inherent in previous fingerprint-mediated methods, achieving new state-of-the-art accuracy on standard mass spectrometry benchmarks without relying solely on reference libraries. The work demonstrates that efficient autoregressive generation can significantly improve recall for identifying molecular structures directly from noisy spectral data.
MS-GPT: Rethinking MS/MS De Novo Structure Elucidation as Spectrum-Induced Posterior Querying of a Molecule-Language Model
arXiv:2607.23607v1 Announce Type: cross 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.
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 this https URL.
18 pages, 14 figures, and 9 tables, including appendices. Source code and model checkpoints are available at this https URL
Machine Learning (cs.LG); Computation and Language (cs.CL)
arXiv:2607.23607 [cs.LG]
(or arXiv:2607.23607v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.23607 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 347 words
Related coverage
MM-Spectrum: Multimodal Multi-spectral Molecular Structural Elucidation with a Stable MoE Framework
arXiv · Aug 28, 2026
Why do models task game?
AI Alignment Forum · Aug 6, 2026
FRIGID: Scaling Diffusion-Based Molecular Generation from Mass Spectra at Training and Inference Time
arXiv · Jul 28, 2026
CoRe-Gen: Robust Spectrum-to-Structure Generation under Imperfect Fingerprint Conditions
arXiv · Sep 30, 2026