Prediction of intercellular communication networks using CellComm

Abstract Intercellular communication is important for tissue development and homeostasis, and when dysregulated contributes to a multitude of pathobiological processes. Cells communicate with each other by several mechanisms, including direct cell-cell contacts between membrane-tethered ligands and receptors on the cell surface, through secreted molecules that bind their cognate receptor on the receiving cell, or alternative modalities such as exosomes. The study of cell-cell communication networks using single-cell genomic approaches is now under intensive investigation, and innovative algorithms to interpret the data, infer how cells interact and identify the downstream effectors of a putative binding of a ligand to its cognate receptor are critically needed. Here, we describe a protocol to run CellComm, a data-driven systems biology algorithm that integrates single-cell RNA-sequencing, protein interaction networks, and gene regulatory networks to infer which cells within a heterogeneous tissue are actively communicating, as well as their downstream transcriptional programs. When spatial transcriptomics data is available, CellComm additionally identifies spatially-resolved cell-cell interactions within the tissue. This protocol is associated with our Nature Cell Biology paper describing our algorithm: CellComm infers cellular crosstalk that drives hematopoietic stem and progenitor cell development.