SpaceWalker: a Visual Analytics Approach to Spatial Transcriptomics Data

Spatial transcriptomics (ST) enables profiling the expression of hundreds of genes in tissue sections, down to the level of single cells in their tissue environment. For single cells, these high-dimensional (HD) gene expression profiles enable detailed characterization of cell types, cell states, and cell maturation. The spatial cell context enables the study of cell-cell interactions, tissue architecture, and cell development and migration in the tissue. Various computational approaches have been developed to extract information from either spatial domain or gene expression domain individually. However, integrative biological interpretation of HD single cell and spatial data spaces remains challenging. The relationship between HD single-cell data, spatial location and similarity embedding has not been fully explored. In this work, we present SpaceWalker, an interactive visual analytics tool for exploring the spatial structure of ST data, while linking it to (developmental) cell phenotype information computed from the HD gene expression profiles. Specifically, we explored approaches where the user is guided by the local intrinsic dimensionality of the HD data to define seed locations for series of random walks; These random walks on the HD KNN graph are then visualized on 2D scatter plots, enabling the user to interactively query for patterns related to cell migration (in the spatial domain) as well as cell maturation (in the HD gene expression domain).

Resources

• Publisher
• Preprint
• Poster

Citation

BibTeX

@misc{ bib:2022_vcbm_chang,
    author = {Chang Li and Thomas H{\"o}llt and Boudewijn Lelieveldt},
    title = { SpaceWalker: a Visual Analytics Approach to Spatial Transcriptomics Data },
    howpublished = { Poster presentation at Poster Presentation, Visual Computing in Biology and Medicine },
    year = { 2022 },
}