Cell-cell interaction

Spateo features numerous methods for investigating cell-cell communication from single-cell spatial transcriptomics, including

  1. Ligand-receptor interaction prediction (by computation of the product) conditioned on spatial proximity (and ligand-receptor enrichment analysis, in the same vein as differential expression to compute spatial region-specific enrichment). This provides an indication of the most prominent pairs of ligands & receptors expressed by cell type pairs, which may be involved in signaling between cells of the given types. For this tutorial, we demonstrate on a 2D example, using a slice collected from the injured telencephalon of an axolotl.

  1. Estimation of the impact of ligand/receptor signaling on expression of genes of interest via spatially-weighted generalized linear modeling. Spateo integrates ligand-receptor interaction predictions with robustly built biological knowledge networks to best infer how ligand-receptor interaction may result in upregulation of particular genes. For this tutorial, we demonstrate on a 3D example; specifically, a portion of the developing diencephalon (in the brain) of a mouse embryo at the E11.5 stage. It is broken into two parts: the first part involves downloading the data and selecting target genes for the model, and the second involves running the model and performing selected downstream analyses.