Advances in Spatial Transcriptomics: Unlocking the Mouse Brain with CellTransformer

Strategic Insights -

Researchers at the University of California have developed a novel workflow, CellTransformer, to analyze organ-scale spatial transcriptomic data. By leveraging a self-supervised framework and encoder-decoder architecture, CellTransformer can learn latent representations of tissue spatial domains and scale to multimillion-cell datasets. This breakthrough enables the detection of fine-grained tissue domains, which can lead to a deeper understanding of the mouse brain's complex structures. The research has been supported by several prominent organizations, including the U.S. Department of Health & Human Services, the Sandler Foundation, and the Weill Neurohub.

Key Takeaways:

  • CellTransformer is a self-supervised framework for learning latent representations of tissue spatial domains, which can be applied to organ-scale spatial transcriptomic data.
  • The workflow uses an encoder-decoder architecture, named CellTransformer, to hierarchically learn higher-order tissue features from lower-level cellular and molecular statistical patterns.
  • CellTransformer can integrate cells across tissue sections, identify domains highly similar to existing ontologies, and discover hundreds of uncataloged areas with minimal loss of domain spatial coherence.
  • The workflow can be applied to multi-million cell MERFISH datasets and whole-brain Slide-seqV2 datasets, allowing for complex multi-animal analyses.
  • CellTransformer achieved nearly perfect consistency of up to 100 spatial domains in a dataset of four individual mice with nine million cells across more than 200 tissue sections.

Statistics:

  • The research utilized a dataset of four individual mice with nine million cells across more than 200 tissue sections.
  • The workflow achieved nearly perfect consistency of up to 100 spatial domains in this dataset.
  • CellTransformer can scale to multimillion-cell datasets, including MERFISH and Slide-seqV2 datasets.
  • The research was supported by several prominent organizations, including the U.S. Department of Health & Human Services, which provided funding for the research.

Sources:

  • Data-driven fine-grained region discovery in the mouse brain with transformers. Nature Communications, 2025;16(1):8536.
  • NewsRx. New Findings from University of California Describe Advances in Information Technology (Data-driven fine-grained region discovery in the mouse brain with transformers). Information Technology Newsweekly. October 21, 2025; p 470.