PlantCellChat: an R-based toolkit for predicting plant cell-cell communication from single-cell and spatial...

Cell-cell communication (CCC) plays a fundamental role in coordinating plant development, environmental responses, and defense mechanisms. However, compared to animal systems, the study of plant CCC has been limited by the lack of comprehensive databases and analysis tools. Here, we present PlantCellChat, a software package designed to facilitate the study of plant CCC at single-cell resolution. PlantCellChat integrates ligand-receptor interaction data from five model plant species (Arabidopsis thaliana, Oryza sativa, Solanum lycopersicum, Zea mays, and Glycine max) and provides a suite of computational tools for predicting CCC networks using single-cell RNA sequencing (scRNA-seq) and spatial transcriptome data. Additionally, we introduce a deep learning-based model, PlantCellChat hormone-graph convolutional networks (PCC-GCN), which classifies plant proteins into specific hormone receptor categories based on their topological and structural features using GCNs. We demonstrate the utility of PlantCellChat through its application to scRNA-seq and spatial transcriptome datasets from rice and Arabidopsis under salt stress and pathogen infection, respectively. Our results predict stress associated alterations in intercellular communication patterns and highlight candidate ligand-receptor pairs potentially involved in stress responses. PlantCellChat is freely available as an R package (available at https://github.com/mrliuw/PlantCellChat) and provides a powerful resource for the plant research community.