Linear structure unfolding : application to mouse brain in spatial transcriptomics

Spatial transcriptomics enables serial analysis of the expression of hundreds of genes in tissue sections by combining molecular biology and imaging approaches. This technique allows the exploration of large tissue sections (up to 3cm²) at the single-cell resolution, allowing cell-cell interactions and subcellular RNA expression analyses. Regions of interest having particular shapes raise specific computational problems that need to be addressed. In this manuscript, we propose a method for the analysis of elongated curved shape (set of connected pixels) or point cloud spread within a shape (cell centers or RNA molecules). We develop a method to unfold linear structures and analyze its content along its principal axis, in order to study cell or transcript type proportions. The centerline of the shape is automatically computed by combining k-means clustering and salesman problem optimization. Our method allows the analysis of spatial distribution of cells or transcripts along or perpendicular to the centerline.