EUVIMEDEUVIMED entdecken
Crossref

CSU-EP: Contrastive Learning for Unifying Experimental and Predicted EI-MS Spectra

Ting Xie, Yaoyu Zhuang, Jiongyu Wang, Hailiang Zhang, Chuanxiu Yu, Zhimin Zhang iD, Hongmei Lu iD

Analytical Chemistry · 04.08.2026 · 98 (30) · 22119-22131

DOI: 10.1021/acs.analchem.6c01233 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Analytical Chemistry, online erschienen am 20.07.2026. Verfasst von Ting Xie und 6 weiteren Autor:innen. Beteiligte Einrichtung: College of Chemistry and Chemical Engineering, Central South University, Changsha410083, China. Die Arbeit nennt 46 Literaturangaben. Gefördert von: National Science and Technology Major Project, Natural Science Foundation of Hunan Province in China, National Natural Science Foundation of China (NSFC).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Spectral library matching is the most common method for compound identification in mass spectrometry (MS), but it is limited by the coverage of experimental libraries. In-silico libraries can expand the searchable chemical space, yet their utility is hindered by the experiment-to-prediction gap arising from the discrepancies between experimental and predicted spectra. Here, we propose CSU-EP, a Contrastive spectral unification framework that bridges this gap with a two-stage training strategy. First, it is self-supervisedly pretrained on 1,883,697 predicted electron ionization (EI) mass spectra using masked peak prediction. Second, it is fine-tuned via contrastive learning to unify representations of paired experimental and predicted spectra. Using this fine-tuned encoder, we assemble 2.24 million NEIMS-predicted spectra into a spectral embedding database (CSU-EP-DB). CSU-EP achieves a Recall@1 of 47.10% on the NIST replib benchmark when searching against CSU-EP-DB, outperforming LLM4MS by 7.28% and FastEI by 12.47%. It also demonstrates excellent ability to identify compounds absent from experimental libraries. In a plasma metabolomics application, incorporating a molecular weight filter boosts Recall@5 to 92.86%. To ensure broad accessibility, a user-friendly web server is deployed. By effectively bridging the experiment-to-prediction gap, CSU-EP establishes a new paradigm for reliable, scalable compound identification using in-silico libraries. The source code, models, and CSU-EP web server are accessible at https://github.com/tingxiecsu/CSU-EP.

Autor:innen und Institutionen

  • College of Chemistry and Chemical Engineering, Central South University, Changsha410083, China.

ORCID-Profile: Zhimin Zhang, Hongmei Lu

Förderung

  • National Science and Technology Major Project
  • Natural Science Foundation of Hunan Province in China
  • National Natural Science Foundation of China (NSFC)

Zitieren

APA

Xie, T., Zhuang, Y., Wang, J., Zhang, H., Yu, C., Zhang, Z., & Lu, H. (2026). CSU-EP: Contrastive Learning for Unifying Experimental and Predicted EI-MS Spectra. Analytical Chemistry, 98(30), 22119-22131. https://doi.org/10.1021/acs.analchem.6c01233

Vancouver

Xie T, Zhuang Y, Wang J, Zhang H, Yu C, Zhang Z, et al. CSU-EP: Contrastive Learning for Unifying Experimental and Predicted EI-MS Spectra. Analytical Chemistry. 2026;98(30):22119-22131. doi:10.1021/acs.analchem.6c01233

BibTeX
@article{Xie2026,
  author = {Xie, Ting and Zhuang, Yaoyu and Wang, Jiongyu and Zhang, Hailiang and Yu, Chuanxiu and Zhang, Zhimin and Lu, Hongmei},
  title = {CSU-EP: Contrastive Learning for Unifying Experimental and Predicted EI-MS Spectra},
  journal = {Analytical Chemistry},
  year = {2026},
  volume = {98},
  number = {30},
  pages = {22119--22131},
  doi = {10.1021/acs.analchem.6c01233}
}
Kostenlos starten →