Advanced search
1 file | 6.26 MB Add to list

Synthetic whole-slide image tile generation with gene expression profile-infused deep generative models

Author
Organization
Abstract
In this work, we propose an approach to generate whole-slide image (WSI) tiles by using deep generative models infused with matched gene expression profiles. First, we train a variational autoencoder (VAE) that learns a latent, lower-dimensional representation of multi-tissue gene expression profiles. Then, we use this representation to infuse generative adversarial networks (GANs) that generate lung and brain cortex tissue tiles, resulting in a new model that we call RNA-GAN. Tiles generated by RNA-GAN were preferred by expert pathologists compared with tiles generated using traditional GANs, and in addition, RNA-GAN needs fewer training epochs to generate high-quality tiles. Finally, RNA-GAN was able to generalize to gene expression profiles outside of the training set, showing imputation capabilities. A web-based quiz is available for users to play a game distinguishing real and synthetic tiles: https://rna-gan.stanford.edu/, and the code for RNA-GAN is available here: https://github.com/gevaertlab/RNA-GAN.
Keywords
Computer Science Applications, Radiology, Nuclear Medicine and imaging, Genetics, Biochemistry, Genetics and Molecular Biology (miscellaneous), Biochemistry, Biotechnology, TUMOR MICROENVIRONMENT, REPRESENTATION, SELECTION, NETWORKS, GENOMICS

Downloads

  • Publisher version.pdf
    • full text (Published version)
    • |
    • open access
    • |
    • PDF
    • |
    • 6.26 MB

Citation

Please use this url to cite or link to this publication:

MLA
Carrillo-Perez, Francisco, et al. “Synthetic Whole-Slide Image Tile Generation with Gene Expression Profile-Infused Deep Generative Models.” CELL REPORTS METHODS, vol. 3, no. 8, Elsevier, 2023, doi:10.1016/j.crmeth.2023.100534.
APA
Carrillo-Perez, F., Pizurica, M., Ozawa, M. G., Vogel, H., West, R. B., Kong, C. S., … Gevaert, O. (2023). Synthetic whole-slide image tile generation with gene expression profile-infused deep generative models. CELL REPORTS METHODS, 3(8). https://doi.org/10.1016/j.crmeth.2023.100534
Chicago author-date
Carrillo-Perez, Francisco, Marija Pizurica, Michael G. Ozawa, Hannes Vogel, Robert B. West, Christina S. Kong, Luis Javier Herrera, Jeanne Shen, and Olivier Gevaert. 2023. “Synthetic Whole-Slide Image Tile Generation with Gene Expression Profile-Infused Deep Generative Models.” CELL REPORTS METHODS 3 (8). https://doi.org/10.1016/j.crmeth.2023.100534.
Chicago author-date (all authors)
Carrillo-Perez, Francisco, Marija Pizurica, Michael G. Ozawa, Hannes Vogel, Robert B. West, Christina S. Kong, Luis Javier Herrera, Jeanne Shen, and Olivier Gevaert. 2023. “Synthetic Whole-Slide Image Tile Generation with Gene Expression Profile-Infused Deep Generative Models.” CELL REPORTS METHODS 3 (8). doi:10.1016/j.crmeth.2023.100534.
Vancouver
1.
Carrillo-Perez F, Pizurica M, Ozawa MG, Vogel H, West RB, Kong CS, et al. Synthetic whole-slide image tile generation with gene expression profile-infused deep generative models. CELL REPORTS METHODS. 2023;3(8).
IEEE
[1]
F. Carrillo-Perez et al., “Synthetic whole-slide image tile generation with gene expression profile-infused deep generative models,” CELL REPORTS METHODS, vol. 3, no. 8, 2023.
@article{01HAN4ZMWZP7ERTA6CJMFMC52N,
  abstract     = {{In this work, we propose an approach to generate whole-slide image (WSI) tiles by using deep generative models infused with matched gene expression profiles. First, we train a variational autoencoder (VAE) that learns a latent, lower-dimensional representation of multi-tissue gene expression profiles. Then, we use this representation to infuse generative adversarial networks (GANs) that generate lung and brain cortex tissue tiles, resulting in a new model that we call RNA-GAN. Tiles generated by RNA-GAN were preferred by expert pathologists compared with tiles generated using traditional GANs, and in addition, RNA-GAN needs fewer training epochs to generate high-quality tiles. Finally, RNA-GAN was able to generalize to gene expression profiles outside of the training set, showing imputation capabilities. A web-based quiz is available for users to play a game distinguishing real and synthetic tiles: https://rna-gan.stanford.edu/, and the code for RNA-GAN is available here: https://github.com/gevaertlab/RNA-GAN.}},
  articleno    = {{100534}},
  author       = {{Carrillo-Perez, Francisco and Pizurica, Marija and Ozawa, Michael G. and Vogel, Hannes and West, Robert B. and Kong, Christina S. and Herrera, Luis Javier and Shen, Jeanne and Gevaert, Olivier}},
  issn         = {{2667-2375}},
  journal      = {{CELL REPORTS METHODS}},
  keywords     = {{Computer Science Applications,Radiology, Nuclear Medicine and imaging,Genetics,Biochemistry, Genetics and Molecular Biology (miscellaneous),Biochemistry,Biotechnology,TUMOR MICROENVIRONMENT,REPRESENTATION,SELECTION,NETWORKS,GENOMICS}},
  language     = {{eng}},
  number       = {{8}},
  pages        = {{17}},
  publisher    = {{Elsevier}},
  title        = {{Synthetic whole-slide image tile generation with gene expression profile-infused deep generative models}},
  url          = {{http://doi.org/10.1016/j.crmeth.2023.100534}},
  volume       = {{3}},
  year         = {{2023}},
}

Altmetric
View in Altmetric