BayesSpace

BayesSpace clusters and enhances the resolution of spatial gene expression experiments by incorporating a spatial prior into a low-dimensional representation of the gene expression matrix to encourage neighboring spots to cluster together.

Key Features:

• Spatially-informed clustering: Clusters a low-dimensional representation of the gene expression matrix with a spatial prior that encourages neighboring spots to cluster together.
• Resolution enhancement: Enhances the resolution of the low-dimensional representation into "sub-spots".
• Feature imputation: Imputes features such as gene expression or cell type composition for the enhanced-resolution sub-spots.

Scientific Applications:

• Spatial transcriptome clustering: Groups spatial gene expression spots into clusters based on expression similarity and spatial proximity.
• Resolution enhancement of spatial transcriptomes: Generates higher-resolution sub-spots from lower-resolution spatial gene expression data.
• Gene expression imputation: Predicts gene expression levels at enhanced-resolution sub-spots.
• Cell type composition imputation: Predicts cell type composition at enhanced-resolution sub-spots.

Methodology:

BayesSpace clusters a low-dimensional representation of the gene expression matrix with a spatial prior to encourage neighboring spots to cluster together, then enhances resolution into sub-spots for feature imputation.