arXiv cs.CL· Xin Zhao, Yumin Liu, Zhuo Li, Weichu Zheng, Feng Zhu, Xiaokang Yang, Yaohui Jin, Yanyan Xu·· 11 小时前AI 评分31
MS-GPT：将 MS/MS 从头结构解析重构为分子语言模型的谱诱导后验查询
MS-GPT: Rethinking MS/MS De Novo Structure Elucidation as Spectrum-Induced Posterior Querying of a Molecule-Language Model
AI 导读
MS-GPT 将指纹介导的 MS/MS 从头结构解析重构为对条件分子语言模型的谱诱导后验查询，通过活性位密度校准把谱诱导后验转为指纹查询带，并按生成频率与后验相似度对候选池排序。
正文 · 原文
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 and posterior similarity. 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 17.1%/27.6% and 10.2%/13.2%, respectively, and candidate-pool scaling improves recovery with a near-linear increase in inference cost. The source code and model checkpoints are available at this https URL.
| Comments: | 18 pages, 13 figures, and 11 tables, including appendices. Source code and model checkpoints are available at this https URL |
| Subjects: | Machine Learning (cs.LG); Computation and Language (cs.CL) |
| Cite as: | arXiv:2607.23607 [cs.LG] |
| (or arXiv:2607.23607v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2607.23607 arXiv-issued DOI via DataCite |
Submission history
From: Xin Zhao [view email]
[v1]
Sun, 26 Jul 2026 11:29:23 UTC (1,698 KB)
[v2]
Thu, 29 Oct 2026 18:04:49 UTC (1,760 KB)
来源：arXiv cs.CL · arxiv.org