A synthetic MALDI-TOF MS atlas for microbial classification across hospitals

Abstract

Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) mass spectrometry (MS) is central to routine clinical microbiology for species identification. Its combination with machine learning is constrained by fragmented public datasets, species imbalance and acquisition-related variability. Synthetic datasets could provide more training data, but they must preserve the spectral information needed to make predictions on real samples. Here, we present an open, synthetic MALDI-TOF MS atlas comprising 4.3 million spectra across 43 clinically relevant bacterial species. To construct the atlas, we harmonised public repositories from three countries and trained a species- and institution-conditioned concept-bottleneck generative variational autoencoder (CBGM-VAE) on 179,086 measured spectra. We assessed the atlas as both a stand-alone training resource and a complement to real data, using bacterial species classification as a functional benchmark. Evaluation comprised 7,269 real spectra from more than 7 countries, 24 institutions and multiple acquisition platforms, covering: temporal, technical, biological and combined shifts. A neural network trained only on synthetic spectra achieved F1 scores ranging from 0.788 to 1.000 across external cohorts. When synthetic spectra were combined with real training data, classification performance improved or matched using real data only. These findings demonstrate that the atlas retains transferable information and can complement measured data. The openly released atlas provides a reusable resource for supervised learning and benchmarking. It can also be explored for pretraining representation-learning and foundation models in MALDI-TOF MS research.

Deciphering the nitrogen-promoting effect of vermicompost: from a complex organic matrix to synthetic microbial communities
Yangfei Li, Jing Huang, Yuanke Liang
Frontiers in plant science16 Oct 2026Peer-reviewedCC BYFull textPlant biologyEurope PMC
Synthetic microbial communities for engineering climate‐smart biofertilizers
Yan Liu, Yuehua Chen, Yanxue Yu +7
iMeta1 Jun 2026Peer-reviewedCC BY7 citedPlant biologyOpenAlex
Genome-scale metabolic models for a synthetic soil microbial community as a path for understanding community functioning
O. Keshk, E. Vayena, S. Causevic +4
bioRxiv16 Oct 2026PreprintCC BY-NC-NDFull textSystems & synthetic biologybioRxiv
Synthetic microbial community promotes seedling growth of Chinese fir via dissolving phosphorus and modifying rhizosphere microbial community
Junhao Zhang, Songxie Xu, Chaoying Chen +5
Tree Physiology21 Sep 2026Peer-reviewedCC BY 4.0Full textPlant biologyPMC
Coupled enzyme discovery, evolution and synthetic yeast chassis adaptation for microbial biopolymer valorisation
M. d. C. Sanchez Olmos, J. Gonzales Isa, N. Paczia +5
bioRxiv28 Sep 2026PreprintCC BY-NCFull textSystems & synthetic biologybioRxiv

Cite

BibTeX RIS

BibTeX

@misc{guerrerolopez2026synthetic,
  title = {{A synthetic MALDI-TOF MS atlas for microbial classification across hospitals}},
  author = {A. Guerrero-Lopez and E. Meijer and Y. Gütlin and K. Koch and N. Rabbani and E. Hitz and E. Serrano-Conde and B. P. Howden and K. K. Sogaard and M. A. Syed and M. Oberle and J. Moran-Gilad and M. Hernandez-Garcia and N. L. Sherry and A. Egli},
  howpublished = {bioRxiv},
  year = {2026},
  month = oct,
  doi = {10.64898/2026.10.19.759651},
  url = {https://www.biorxiv.org/content/10.64898/2026.10.19.759651v1},
  note = {CC BY},
}

RIS

TY  - UNPB
TI  - A synthetic MALDI-TOF MS atlas for microbial classification across hospitals
AU  - A. Guerrero-Lopez
AU  - E. Meijer
AU  - Y. Gütlin
AU  - K. Koch
AU  - N. Rabbani
AU  - E. Hitz
AU  - E. Serrano-Conde
AU  - B. P. Howden
AU  - K. K. Sogaard
AU  - M. A. Syed
AU  - M. Oberle
AU  - J. Moran-Gilad
AU  - M. Hernandez-Garcia
AU  - N. L. Sherry
AU  - A. Egli
PY  - 2026
DA  - 2026/10/19
PB  - bioRxiv
DO  - 10.64898/2026.10.19.759651
UR  - https://www.biorxiv.org/content/10.64898/2026.10.19.759651v1
AB  - Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) mass spectrometry (MS) is central to routine clinical microbiology for species identification. Its combination with machine learning is constrained by fragmented public datasets, species imbalance and acquisition-related variability. Synthetic datasets could provide more training data, but they must preserve the spectral information needed to make predictions on real samples. Here, we present an open, synthetic MALDI-TOF MS atlas comprising 4.3 million spectra across 43 clinically relevant bacterial species. To construct the atlas, we harmonised public repositories from three countries and trained a species- and institution-conditioned concept-bottleneck generative variational autoencoder (CBGM-VAE) on 179,086 measured spectra. We assessed the atlas as both a stand-alone training resource and a complement to real data, using bacterial species classification as a functional benchmark. Evaluation comprised 7,269 real spectra from more than 7 countries, 24 institutions and multiple acquisition platforms, covering: temporal, technical, biological and combined shifts. A neural network trained only on synthetic spectra achieved F1 scores ranging from 0.788 to 1.000 across external cohorts. When synthetic spectra were combined with real training data, classification performance improved or matched using real data only. These findings demonstrate that the atlas retains transferable information and can complement measured data. The openly released atlas provides a reusable resource for supervised learning and benchmarking. It can also be explored for pretraining representation-learning and foundation models in MALDI-TOF MS research.
ER  -

Abstract

Related

Cite