Characterizing tissue structures from spatial omics with spatial cellular graph partition

Abstract

Spatial transcriptomic and proteomic measurements enable high-dimensional characterization of tissues. However, understanding organizations of cells at different spatial scales and extracting tissue structures of interest remain challenging tasks that require extensive human annotations. To address this need for consistent identification of tissue structures, in this work, we present a novel annotation method Spatial Cellular Graph Partitioning (SCGP) that allows unsupervised identification of tissue structures that reflect the anatomical and functional units of human tissues. We further present a reference-query extension pipeline SCGP-Extension that enables the generalization of existing reference tissue structures to previously unseen samples. Our experiments demonstrate reliable and robust partitionings of both spatial transcriptomics and proteomics datasets encompassing different tissue types and profiling techniques. Downstream analysis on SCGP-identified tissue structures reveals disease-relevant insights regarding diabetic kidney disease and skin disorder, underscoring its potential in facilitating spatial analysis and driving new discoveries.