Skip to content
DraconiticResearch
Search Sources API
186,625 studies31 sourcesUpdated 18 min ago
Peer-reviewedCC BYAnalytical Chemistry

SpecTUS: Spectral Translator for Unknown Structures Annotation from EI-MS Spectra

  • Adam Hájek1
  • Michal Starý2
  • Elliott James Price3
  • Filip Jozefov4
  • Helge Hecht3
  • Aleš Křenek1
4 affiliations
  1. 1Masaryk University , , Šumavská 525/33 , ,
  2. 2Technical University Munich , , Boltzmannstraße 3 , ,
  3. 3Masaryk University , , Kamenice 753/5 , ,
  4. 4Masaryk University , , Botanická 554/68a , ,
Open at sourcePDFDOICite

Abstract

Compound identification and structure annotation from mass spectra are essential in drug detection, forensics, and small molecule discovery. Current approaches to compound identification from electron ionization mass spectra (EI-MS) are dependent on different forms of searching databases that are orders of magnitude smaller than the space of potential molecular structures they attempt to cover. We introduce SpecTUS: Spectral Translator for Unknown Structures, a deep learning model for de novo structural annotation, translating gas chromatography EI-MS spectra directly into molecular structures without requiring reference databases. This enables the identification of novel compounds absent from spectral libraries. In a rigorous evaluation, SpecTUS significantly outperformed standard database search techniques. On a held-out test set of 28,267 spectra from NIST 20, the model's single suggestion perfectly reconstructed 43% of the subset's compounds. On 76% of this test set, the single suggestion is strictly better, in terms of Tanimoto similarity of Morgan fingerprint, than the result of hybrid database search. With ten suggestions, SpecTUS achieved 65% perfect reconstructions, surpassing hybrid search on 84% of the test set.

Related

  1. Closed-loop formulation–process optimization under unknown feasibility for extended-release pellets
    Pauric Bannigan, Nico Saputra, Michael Craig +2
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv
  2. AriadneX: Unknown Crystalline Phase Discovery from Anomaly-Aware Unmixing of Powder X-Ray Diffraction Patterns
    Zhuoran Yang, Christopher M. Collins, Luke M. Daniels +5
    ChemRxivPreprintCC BYFull textMaterials & polymersChemRxiv
  3. A transferable perturbation proteomics framework for phenotype- 1 based annotation and antiproliferative potency scoring of antineoplastic agents
    Jing Liang, Jundan Tian, Rui Zhang +10
    ChemRxivPreprintCC BY-NC-NDPhysical chemistryChemRxiv
  4. Centerline-Guided Weld Annotation for PCCP: Methods and Application Performance
    Zhenyu Chen, Kuanmin Mao, Junjie Tang +1
    Applied SciencesPeer-reviewedCC BYFull textGeneral biologyGeneral chemistryDOAJ
  5. Serum per- and polyfluoroalkyl substances levels among young adults at risk for chronic kidney disease of unknown etiology in Nicaragua
    Leanne S. Fawkes, Yau Adamu, David Gimeno Ruiz de Porras +3
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textEnvironmental chemistryPMC

Cite

BibTeXRIS
BibTeX
@article{hajek2026spectus,
  title = {{SpecTUS: Spectral Translator for Unknown Structures Annotation from EI-MS Spectra}},
  author = {Adam Hájek and Michal Starý and Elliott James Price and Filip Jozefov and Helge Hecht and Aleš Křenek},
  journal = {Analytical Chemistry},
  year = {2026},
  month = jul,
  doi = {10.1021/acs.analchem.6c02423},
  url = {https://doi.org/10.1021/acs.analchem.6c02423},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - SpecTUS: Spectral Translator for Unknown Structures Annotation from EI-MS Spectra
AU  - Adam Hájek
AU  - Michal Starý
AU  - Elliott James Price
AU  - Filip Jozefov
AU  - Helge Hecht
AU  - Aleš Křenek
PY  - 2026
DA  - 2026/07/13
JO  - Analytical Chemistry
DO  - 10.1021/acs.analchem.6c02423
UR  - https://doi.org/10.1021/acs.analchem.6c02423
AB  - Compound identification and structure annotation from mass spectra are essential in drug detection, forensics, and small molecule discovery. Current approaches to compound identification from electron ionization mass spectra (EI-MS) are dependent on different forms of searching databases that are orders of magnitude smaller than the space of potential molecular structures they attempt to cover. We introduce SpecTUS: Spectral Translator for Unknown Structures, a deep learning model for de novo structural annotation, translating gas chromatography EI-MS spectra directly into molecular structures without requiring reference databases. This enables the identification of novel compounds absent from spectral libraries. In a rigorous evaluation, SpecTUS significantly outperformed standard database search techniques. On a held-out test set of 28,267 spectra from NIST 20, the model's single suggestion perfectly reconstructed 43% of the subset's compounds. On 76% of this test set, the single suggestion is strictly better, in terms of Tanimoto similarity of Morgan fingerprint, than the result of hybrid database search. With ten suggestions, SpecTUS achieved 65% perfect reconstructions, surpassing hybrid search on 84% of the test set.
SN  - 0003-2700
SN  - 1520-6882
ER  -