Abstract
ABSTRACTA wide range of pathologies and chronic diseases are characterized by transformations in cell type composition within a tissue. Prominent examples are remodeling of barrier epithelia in response to external threats, immune cell infiltration in chronic inflammatory conditions, or proliferation of fibroblasts and stromal cells causing fibrosis. An extensive literature implicates paracrine signals secreted by the local tissue niche in regulating its cell type composition, and therefore dysregulation of these factors is frequently capable of inducing pathological changes. Emerging multi-omics methods, particularly spatial transcriptomic profiling, provide the ability to simultaneously assay cell type composition and gene expression, with the potential to discover the transcripts most strongly associated with compositional shifts. However, no method yet exists to detect such signals. Here, we develop Spatial Paired Expression Ratio (SPER), a computational approach to evaluate the spatial dependence between transcripts and cell types in spatial transcriptomics data. We demonstrate the ability of SPER to accurately detect paracrine drivers of cellular abundance using simulated datasets. Using publicly available spatial transcriptomic data from the adult mouse motor cortex, we show that genes identified by SPER are strongly statistically enriched for both transcripts known to be extracellularly secreted, and those known to participate in paracrine receptor-ligand pairs, providing evidence of the ability to detect compositional regulatory signals. SPER recovered known and novel interactions, such as a surprising spatial association between the Wnt ligandRpso3and its cognate receptorLgr5, expressed in a specific neural subset, and not previously known to regulate neural cell type composition in the motor cortex. SPER thus presents a general approach to discovering paracrine drivers of cellular compositional changes.
Publisher
Cold Spring Harbor Laboratory