Screening cell-cell communication in spatial transcriptomics via collective optimal transport

Abstract

Spatial transcriptomic technologies and spatially annotated single cell RNA-sequencing (scRNA-seq) datasets provide unprecedented opportunities to dissect cell-cell communication (CCC). How to incorporate the spatial information and complex biochemical processes in reconstructing CCC remains a major challenge. Here we present COMMOT to infer CCC in spatial transcriptomics, which accounts for the competition among different ligand and receptor species as well as spatial distances between cells. A novel collective optimal transport method is developed to handle complex molecular interactions and spatial constraints. We introduce downstream analysis tools on spatial directionality of signalings and genes regulated by such signalings using machine learning models. We apply COMMOT to simulation data and eight spatial datasets acquired with five different technologies, showing its effectiveness and robustness in identifying spatial CCC in data with varying spatial resolutions and gene coverages. Finally, COMMOT reveals new CCCs during skin morphogenesis in a case study of human epidermal development. Both the method and the computational package have broad applications in inferring cell-cell interactions within spatial genomics datasets.