GASTON - Mapping the topography of spatial gene expression with interpretable deep learning

GASTON is an interpretable deep learning model for learning the gene expression topography of a tissue slice from spatially resolved transcriptomics (SRT) data. GASTON models gene expression topography by learning the isodepth, a 1-D coordinate describing continuous gene expression gradients and tissue geometry (i.e. spatial domains).

GASTON model architecture

GASTON’s key applications

  • Learns 1-d coordinate that varies smoothly across tissue slice, providing topographic map of gene expression in the tissue slice.

  • Modeling continuous gradients of gene expression for individual genes, e.g. gradients of metabolism in cancer

  • Identifying tissue geometry, i.e. arrangement of spatial domains

GASTON model architecture

Manuscript

Please see our manuscript [] for more details.

Getting started with GASTON

  • Browse Tutorial for a quick start guide to GASTON.

  • Discuss usage and issues on github.