A0881
Title: InterSpatial: Leveraging low-resolution spatial transcriptomics to infer cell-cell communication in scRNA-seq
Authors: Tuhin Majumder - Wake Forest University (United States) [presenting]
Abstract: Cell-cell communication (CCC) is often influenced by spatial proximity, prompting recent methods to incorporate spatial information using spatial transcriptomics (ST) data. However, in single-cell RNA-seq data, spatial coordinates are missing, limiting CCC inference to ligand-receptor expression patterns alone, potentially leading to incomplete or inaccurate conclusions. InterSpatial is introduced, a novel framework that leverages publicly available low-resolution ST data to enhance CCC analysis in single-cell data. InterSpatial performs cellular mapping, meta-cell clustering, and optimal transport to estimate inter-meta-cell distances. It examines ligand expression in sender meta-cells and trends of receptor expression in receiver meta-cells with respect to spatial distance, offering a visualization tool to screen for plausible CCC. Additionally, InterSpatial enables the application of ST-based CCC methods as downstream analyses. Its effectiveness is demonstrated by analyzing CCC between senescence-susceptible cells and macrophages in idiopathic pulmonary fibrosis (IPF) lungs, microglia and neurons in Alzheimer's disease brain tissue, and POSTN-positive fibroblasts and myeloid cells in myocardial infarction (MI) heart tissue. Results highlight the ability of InterSpatial to recover spatial context and uncover novel CCC patterns.
