Modeling Dynamic Cross-Correlation in Spatial Transcriptomic Data

In recent years, developments in high-throughput spatial transcriptomics technologies have made it possible to comprehensively map gene expression with fine-grain resolution across entire tissue sections. Leveraging this advanced level of spatial information, researchers have demonstrated that gene expression patterns and signaling pathways in cells are significantly impacted by their spatial proximity to other cell types, such as t-cells, cancer-associated-fibroblasts (CAFs), and tumor cells, as well as by their spatial location within a cluster of their own cell type. A number of methods have been proposed in the literature for using spatial transcriptomics data to analyze how gene expression varies dependent on a cell’s spatial location, but much less work has been done to analyze how the intra- and intercellular correlation between genes varies over space.

Journal article