EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Learning interpretable representations of single-cell multi-omics data with multi-output Gaussian processes

Zahra Moslehi, Sareh AmeriFar, Kevin de Azevedo, Florian Buettner iD

Nucleic Acids Research · 2025

DOI: 10.1093/nar/gkaf630 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Nucleic Acids Research, 2025. Verfasst von Zahra Moslehi und 3 weiteren Autor:innen. Beteiligte Einrichtung: German Cancer Consortium (DKTK), partner site Frankfurt/Mainz, a partnership between DKFZ and UCT Frankfurt-Marburg, 60590 Frankfurt am Main, Germany. Hauptthemen laut MEDLINE: Single-Cell Analysis, Genomics, Machine Learning. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: European Union, ERC, TAIPO.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Learning representations of single-cell genomics data is challenging due to the nonlinear and often multi-modal nature of the data on one hand and the need for interpretable representations on the other hand. Existing approaches tend to focus either on interpretability aspects via linear matrix factorization or on maximizing expressive power via neural network-based embeddings using black-box variational autoencoders or graph embedding approaches. We address this trade-off between expressive power and interpretability by introducing a novel approach that combines highly expressive representation learning via an embedding layer with interpretable multi-output Gaussian processes within a unified framework. In our model, we learn distinct representations for samples (cells) and features (genes) from multi-modal single-cell data. We demonstrate that even a few interpretable latent dimensions can effectively capture the underlying structure of the data. Our model yields interpretable relationships between groups of cells and their associated marker genes: leveraging a gene relevance map, we establish connections between cell clusters (e.g. specific cell types) and feature clusters (e.g. marker genes for those specific cell types) within the learned latent spaces of cells and features.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Single-Cell AnalysisGenomicsMachine LearningNormal DistributionHumansNeural Networks, ComputerAlgorithmsMultiomics

Autor:innen und Institutionen

  • German Cancer Consortium (DKTK), partner site Frankfurt/Mainz, a partnership between DKFZ and UCT Frankfurt-Marburg, 60590 Frankfurt am Main, Germany.
  • German Cancer Research Center (DKFZ), 69120 Heidelberg, Germany.
  • Department of Medicine, Goethe University Frankfurt, 60590 Frankfurt am Main, Germany.
  • Department of Informatics, Goethe University Frankfurt, 60325 Frankfurt am Main, Germany.
  • Frankfurt Cancer Institute (FCI), 60596 Frankfurt am Main, Germany.

ORCID-Profile: Florian Buettner

Förderung

  • European Union
  • ERC
  • TAIPO
  • Deutsche Forschungsgemeinschaft
  • DFG
  • German Research Foundation

Zitieren

APA

Moslehi, Z., AmeriFar, S., de Azevedo, K., & Buettner, F. (2025). Learning interpretable representations of single-cell multi-omics data with multi-output Gaussian processes. Nucleic Acids Research. https://doi.org/10.1093/nar/gkaf630

Vancouver

Moslehi Z, AmeriFar S, de Azevedo K, Buettner F. Learning interpretable representations of single-cell multi-omics data with multi-output Gaussian processes. Nucleic Acids Research. 2025. doi:10.1093/nar/gkaf630

BibTeX
@article{Moslehi2025,
  author = {Moslehi, Zahra and AmeriFar, Sareh and de Azevedo, Kevin and Buettner, Florian},
  title = {Learning interpretable representations of single-cell multi-omics data with multi-output Gaussian processes},
  journal = {Nucleic Acids Research},
  year = {2025},
  doi = {10.1093/nar/gkaf630}
}
Kostenlos starten →