sequencing-based spatial transcriptomics

The analysis of spatial gene expression data generated by the Visium system begins with preprocessing the raw data with space ranger. Briefly, it consists in performing demultiplexing and quality control to filter out low-quality reads, aligning the remaining reads to a reference before counting gene expression from a single capture area. Once the matrix of gene counts per spot has been obtained, the aim is to identify the cell types represented in each spot. Although Visium technology avoids the dissociation step, its resolution is too low to perform single-cell characterization, as each 55µm diameter spot contains several cells. This implies that each spot is a combination of the transcriptome of potentially heterogeneous populations, which can be estimated using deconvolution methods, leveraging or not on a single-cell RNA-seq expression profils reference. After deconvolution, we can compare the distribution of cell types by histological region within the same section, or between sections from different experimental conditions. Combined with the spatial gene expression patterns, spots annotation can be used for Niche analysis, in order to characterize the interactions and relationships between cells and their microenvironment within tissues.